SDSwiki 2017

System Design Strategies
From wiki.GIS.com
System Design Strategies 41st Edition - ArcGIS 10.5.1
An Esri ® Technical Reference Document • Fall 2017

Dave Peters
Manager, Systems Integration
Email: dpeters@esri.com
From wiki.GIS.com
System Design Strategies 41st Edition - ArcGIS 10.5.1
An Esri ® Technical Reference Document • Fall 2017

Dave Peters
Manager, Systems Integration
Email: dpeters@esri.com
Twitter: @DavePetersSDS
Specific license terms for this content

System Design Strategies 26th edition - An Esri ® Technical Reference Document • 2009 (final PDF release)
Table of Contents
Preface (Executive Summary)
1. System Design Process

2. GIS Software Technology
3. Software Performance
4. Server Software Performance
5. GIS Data Administration
6. Network Communications
7. Platform Performance
8. Information Security
9. GIS Product Architecture
10. Performance Management
11. City of Rome
12. System Implementation
Appendixes
A1. Capacity Planning Tool
• Capacity Planning Tool updates
• Capacity Planning Tool Administration
A2. Enterprise System Design Planning Tools
A3. Acronyms and Glossary
B1. Windows Memory Management
B2. System Design Strategies popular pages
C1. CPT Virtualization Model
D1. Planning for Building a GIS video series
System Design Strategies Preface
Fall 2017 System Design Strategies Preface 41st Edition
This System Design Strategies wiki Software Performance

documentation is shared to help Esri This section shares lessons learned about selecting
customers build and maintain successful GIS and building effective GIS design solutions that satisfy
operations. The audience includes GIS operational performance and scalability needs.
managers, Project managers, GIS technical • CPT Software Performance demo
architects, IT administrators, business
partners, systems integrators, system
consultants, system administrators, and Server Software Performance
software developers. The content is focused There are many key factors that impact ArcGIS
Server software performance and scalability. Proper
on system design fundamentals - the
configuration and services deployment can empower
technology concepts that contribute to your GIS operations.
successful GIS operations. • CPT Server Software Performance demo
GIS technology is rapidly evolving to leverage

cloud computing platforms, including a full GIS Data Administration
suite of applications and services to discover, Data provides the resources you need to make proper
business decisions. The information products required
create, manage, visualize, analyze,
to make business decisions determine the critical data
collaborate and share a better understanding resources that must be available for business
of our world. operations. How you organize and maintain your data
resources can determine the success of your
The chapters that follow provide an overview business operations.
of the many pieces of technology that must • GIS Data Administration demo
work together within a distributed GIS
production environment. Online streaming Network Communications
video links are provided at the end of several Network communications provide the connectivity for
chapters to demonstrate how the Capacity distributed GIS operations. Network capacity can limit
Planning Tool can be used to answer your alternative software technology solutions that perform
performance and scalability questions and well within your organization. System architecture
complete your system architecture design. design can identify and address network
communication constraints and identify the right
technical solution for a successful GIS
System Design Process implementation.
System architecture design is an analytic process that • CPT Network Communications demo
generates IT hardware and network capacity
requirements from specified peak business workflow Platform Performance
performance needs. Selecting the right hardware will improve user
• CPT System Design Process demo performance, reduce overall system cost, and
establish a foundation for building effective GIS
operations. Selecting the wrong hardware can
GIS Software Technology contribute to implementation failure - spending money
A variety of ArcGIS software technology patterns are on a system that will not support your business
available to satisfy enterprise business operational needs.
needs. Technology patterns include a broad range of • CPT Platform Performance demo
Desktop, Server, Portal and Mobile deployment
options. Selecting the right technology is a critical
step in the enterprise system design process. Information Security
• CPT GIS Software Technology demo This chapter identifies the purpose and scope of
information security. Basic concepts are introduced
for developing security solutions that satisfy your
business needs.
System Design Strategies Preface 0.1

GIS Product Architecture System Implementation
GIS Product Architecture provides a foundation for
Successful implementation depends on a solid
understanding the software components and platform design, appropriate hardware and software
configuration options available for distributed GIS selection, successful systems integration, and
operations. Understanding application architecture careful evaluation during installation. A
alternatives and associated configuration strategies phased approach to implementation reduces
provides a foundation for selecting an appropriate project risk and promotes success, providing
distributed GIS design.
opportunity for early success and flexibility to
• CPT Calculator architecture demo
incorporate new technology at low risk.
• CPT Design Generic architecture demo
• CPT Design Detailed architecture demo Capacity Planning Tool Appendix
Performance Management The Esri Capacity Planning Tool (CPT) is used
throughout the System Design Strategies (SDS) wiki
Performance management involves building a design to demonstrate design concepts and automate the
solution based on appropriate workflow performance system design analysis. The CPT demonstrations
targets and managing compliance throughout design reinforce design concepts presented throughout the
and implementation to deliver within those targets. SDS wiki chapters.
• CPT Calculator and Design productivity
adjustment capabilities demo The following wiki chapters document the Capacity
• CPT platform performance demo Planning videos.
• CPT Test performance validation demo System Design Process (CPT Demos)
GIS Software Technology (CPT Demos)
City of Rome Software Performance (CPT Demos)
City of Rome is the name of a case study that Server Software Performance (CPT Demos)
demonstrates a process that can be used as a
foundation for most enterprise system design GIS Data Administration (CPT Demos)
analysis. This case study makes use of the Esri Network Communications (CPT Demos)
Capacity Planning Tools for modeling user
requirements and completing the system architecture Platform Performance (CPT Demos)
design GIS Product Architecture (CPT Calculator
Demos)
City of Rome CPT demos
GIS Product Architecture (CPT Design
• City of Rome Executive Summary
Demos)
• City of Rome existing operations
• Phase 1: ArcGIS Enterprise IOC Performance Management (CPT Demos)
• Phase 1 Vector tile basemap deployment City of Rome (CPT Demos)
• Phase 1: Upgrade WTS host platform tier City of Rome Year 1 (CPT Demos)
to support remote ArcGIS Pro clients City of Rome Year 2 (CPT Demos)
• Phase 2: Water Utility Solution
System Design Strategies Preface 0.2

System Design Process
Fall 2017 System Design Process 41st Edition
The purpose of this document is to share a
System architecture design is a process
system design methodology that promotes
developed by Esri to promote successful GIS
successful deployment of geographic
enterprise operations. This process builds on
information system (GIS) technology. This
your existing information technology (IT)
system design methodology includes
infrastructure and provides specific
guidelines for identifying business
recommendations for hardware and network
requirements, making appropriate software
solutions based on existing and projected
selection, using properly configured data
business (user) needs.
sources, and providing sufficient hardware to
meet user productivity needs. This document The system architecture design process aligns
focuses on system performance and identified business requirements (user needs)
scalability - building a GIS that will perform derived from business strategy, goals, and
during peak operational loads. drivers (business processes) with identified
business information systems infrastructure
Much more can be said about business
technology (network and platform)
requirements analysis (GIS User Needs) and
recommendations. The computer display
available software functionality - this is not the
transaction is the work unit used to translate
focus of this documentation. Dr. Roger
business requirements to associated server
Tomlinson (Father of GIS) provides an
and network loads. Display service times
excellent book that shares a proven
(software processing) generates platform
framework for comprehensive GIS planning
loads and display traffic generates network
called Thinking about GIS. Understanding the
loads. Peak throughput loads (business
information products you want out of the GIS
needs) are used to generate platform capacity
and identifying the software candidates you
and network bandwidth requirements.
might use to produce these information
products is a prerequisite for completing your System design starts with identifying business
system architecture design. The Planning for needs. This includes identifying user locations
Building a GIS video series shares an and required information products, identifying
overview of this business process analysis required data resources, and developing
planning methodology. appropriate software applications to do the
work. Business needs are represented by
project workflows that identify the traffic and
What Is System Architecture processing associated with each display
transaction.
Design? System architecture design translates
business needs to identified IT requirements.
Hardware requirements are generated based
on peak software processing loads. Network
connectivity requirements are generated
based on peak data flow. Capacity Planning
tools are provided to automate the design
analysis. Capacity Planning Tools make the
process of aligning Business workflows with
selected IT resources agile and iterative in
nature, rapidly identifying system performance
Figure 1.1 System architecture design is an impacts in response to changing business and
analytic process that generates IT hardware and technology architecture patterns.
network capacity requirements from specified peak
business workflow performance needs.
System Design Process 1.1

Warning: Trying to build a GIS without performance needs and bandwidth constraints
completing a proper system architecture over distributed communication networks—
design can lead to system deployment technology and solution architecture must be
failure. selected to conserve shared infrastructure
resources. System architecture design can
provide a solid foundation for building a
Why Is System Architecture Design productive operational environment.
Important?
Workflow complexity determines processing
and data flow loads that must be handled by
the computing infrastructure. The computing
architecture must be selected with the
appropriate capacity to service the required
business loads. Workflow complexity is a
measure of the amount of processing loads
and network traffic required to refresh the user
display. Complexity is imposed on the
computing architecture by the following design
Figure 1.2 System performance chain. Overall attributes:
system performance is limited by the weakest link • Database design and data format: DBMS,
in the performance chain.
geodatabase, and ArcSDE
System performance is limited by the weakest • User workflow software design:
link in the system design. System architecture Application development
design identifies the weak links during the
planning process and promotes investment in The computing infrastructure must provide
a balanced system design. sufficient capacity to handle peak operational
loads.
A distributed computer environment must be
• Server platform processor core and
designed properly to support user
deployment architecture must handle
performance (productivity) requirements. The
peak processing loads.
weakest "link" in the system will limit
• Network bandwidth and remote site
performance. The system architecture design
connectivity must be adequate to avoid
process identifies specifications for a
traffic contention.
balanced hardware solution. Investment in
• Storage access performance and capacity
hardware and network components based on
must be adequate to provide required
a balanced system load model provides the
data access.
highest possible system performance at the
lowest overall cost as represented by the Server performance, network capacity, and
chain in Figure 1.2. efficient storage strategies can improve user
productivity and reduce system cost.
Building a high-performance GIS requires
more than getting the hardware right. User The system architecture design process can
workflows must be designed to optimize client be used to identify specifications for a
productivity (simple maps) and efficiently balanced hardware solution. Investment in
manage heavier geoprocessing loads (service hardware and network components based on
request queue). The geodatabase design and a balanced system load model provides the
data source selection should be optimized to highest possible system performance at the
address system performance and scalability lowest overall cost.
requirements. The selected production
platform components (servers, workstations,
storage) must have the capacity to handle
peak user workflow processing loads within an
acceptable service response time. The system
architecture design must address

Why we do planning will satisfy identified business requirements is
fundamental to project success.
Understanding key system performance
parameters, identifying incremental system
performance targets, and establishing a
system performance validation plan can chart
a solid framework for managing
implementation risk.
Planning is an incremental process driven by
rapidly changing technology. The most
effective plans are designed to adapt to
changing business needs and evolving
Figure 1.3 Planning is required to quantify business technology opportunities. The established
requirements. change control process must be agile and
The primary reasons for planning include effective to ensure project tasks are managed
identifying business needs, defining project to deliver within established schedule and
requirements, and reducing implementation budget constraints.
risk. In practical terms, we need a plan if we
hope to get something done. The plan What questions are we trying to
provides a foundation for successful answer?
implementation. Figure 1.3 identifies some of
the primary planning objectives.
Business requirements may include a variety
of user workflows and Web services that
would improve business operations – these
workflows and services must be identified
early during the planning process. The
workflow and service definitions establish a
foundation for the rest of the planning.
The peak user workflow and service
processing requirements must be identified
and translated to appropriate system design
loads – server processing times and network
traffic. These processing loads provide a
foundation for generating appropriate
Figure 1.4 What questions do you have about
hardware specifications (performance and building your GIS?
capacity requirements) and network
infrastructure needs. There are many questions about system
architecture design that we answer on a daily
A project must be defined to procure and basis. The System Design Strategies
implement the required technology. The documentation deployed on this wiki site is
project will include a budget and schedule developed and maintained to answer most of
along with a variety of project performance these questions, and provide a reference to
milestones. Project approval is required help customers better understand their design
before any major design and/or needs.
implementation efforts begin. Project approval
is based on identified or perceived benefits Here is a short list of some of the more
that will accrue to the business resulting from common questions:
the proposed level of time (schedule) and • How many users can I support with my
financial (budget) investment. existing hardware?
A good plan can be used to reduce system • What hardware do I need to purchase?
implementation risk. Delivering a system that • How many servers (cores) do I need?

• What are the software licensing • Processing: Platform processor core
requirements? processing demands
• What workflow loads should I use for my • Memory: Platform physical memory
existing applications? utilization
• What are my current workflow service • Graphics: Video graphics card loads
times? • Disk I/O: Storage disk access
• What is the capacity of my current performance
system?
Different workflows place different loads on
the system. Relative processing demands for
This documentation, along with the the four use patterns above show how these
companion Capacity Planning Tools, was loads are distributed differently, based on the
developed so you can answer these questions workflow.
about your GIS, and many more. We have
• Data query and analysis is network I/O
been supporting GIS customers for over 25
intensive.
years, and much of what we have learned is
• Analysis and processing is compute
shared so you can benefit and build a
(processing) intensive.
successful GIS.
• Fly-through and 3D animation can be
graphics intensive.
What demands does GIS place on • Data loading and conversion can be disk
the computing infrastructure? I/O intensive.
GIS operations are typically network and/or
processing intensive—these are the
subsystems that require the closest attention.
GIS applications place heavy demands on
server processing and network bandwidth
resources. For most standard mapping
operations, these demands push the limits of
platform processor and network bandwidth
capacity.
Additional critical design components include
the following:
• Memory is critical if you do not have
enough; memory requirements can
normally be satisfied by following
standard configuration guidelines.
Platform memory guidelines are provided
Figure 1.5 These charts show relative performance
in Windows Memory
demands on the basic infrastructure components,
and how these performance demands vary with
Management appendix.
four very different processing scenarios. • Some GIS desktop applications are
graphics intensive (3D animation); video
GIS applications place heavy demands on graphics are less critical for GIS than for
server processing and network bandwidth the video gaming industry, so there is
resources. For most standard mapping plenty of graphics technology available at
operations, these demands push the limits of reasonable cost to meet GIS needs.
platform processor (CPU) and network • Disk I/O can be a concern if you are
bandwidth (I/O) capacity. moving large volumes of data in or out of
Several key infrastructure components work storage, such as with imagery or map
together to service business processing loads. cache generation (normally a background
administrative process).
• Network I/O: Network traffic demands on
available bandwidth

Any one of these components, if configured
incorrectly, can become a bottleneck and limit
What is the System Design
system performance. Process?
System architecture design is about building a The system design process includes a GIS
balanced solution that satisfies peak needs assessment and a system architecture
performance needs at the lowest possible design. The system architecture design is
cost. Each of these critical subsystems should based on user workflow requirements
be considered in establishing the right system identified in the GIS needs assessment.
design.
Traditional system design process
Cost of a change GIS Needs Assessment: The GIS needs
assessment includes a review of user
workflow requirements and identifies where
GIS technology can improve the quality and
productivity of the business process flow. This
assessment identifies GIS application and
data requirements and an implementation
strategy for supporting GIS user needs. The
user organization must be actively involved
throughout the user needs assessment. A GIS
solutions architect familiar with current GIS
technology patterns and customer business
practices can help facilitate this planning
effort.
Figure 1.6 The cost of making a change increases
rapidly as you purchase and apply GIS System Architecture Design: The system
implementation resources. architecture design is based on user
Getting it right from the start is best done by requirements identified by the GIS needs
taking the time to understand the technology, assessment. Customers must have a clear
quantify user requirements, select the right understanding of their GIS application and
software technology, and deploy the right data requirements before they are ready to
hardware. Not getting it right from the start will complete their system architecture design. An
cost money to make it right later. The cost of enterprise architect familiar with GIS software
change increases exponentially as patterns and system design analysis is
implementation proceeds as shown in Figure typically used to complete the final system
1.6. design and analysis. Technology purchase
decisions should be delayed until after
Your GIS plan should include performance completing the system architecture design
validation early in the design process. ArcGIS analysis.
for Desktop authors should have performance
targets they verify when publishing a new map The system design begins with a technology
service. GIS programmers should have exchange. The technology exchange provides
performance targets they build to when a foundation for client support throughout the
developing a new application. Performance design process. Client participation is a key
should be a consideration for data ingredient in the design process. The design
administrators, network administrators, and process includes a review of the existing
selected storage subsystems. Getting the computer environment, GIS user
design right from the start can save money requirements, and system design alternatives.
and improve overall business operations. System design capacity planning tools can be
used to translate projected peak user
workflow requirements to specific platform
specifications. An integrated implementation

strategy can then be developed to support > The Application Architecture
GIS deployment milestones. provides a blueprint for the individual
applications to be deployed, their
Warning: Problems can occur when
interactions, and their relationships to
software technology selection is made
the core business processes of the
based on a GIS needs assessment without
organization.
considering system architecture design
• The Technology Architecture describes
implications.
the logical software and hardware
Enterprise business needs capabilities that are required to support
assessment the deployment of business, data, and
application services. This includes IT
infrastructure, middleware, networks,
communications, processing, standards,
etc.
A primary goal behind developing capacity

planning tools is to automate the system
architecture design analysis in such a way
that GIS solution architects can use these
tools to complete a user needs assessment
and a system architecture design within an
enterprise level business needs assessment.
Figure 1.7 Enterprise level needs assessment An enterprise level business needs
promotes optimum business integration and assessment promotes an adaptive discovery
investment benefits.
effort, enabling optimum alignment between
Traditionally, the user needs assessment and evolving business and technology architecture
the system architecture design were two patterns.
separate efforts. There are some key
advantages in completing these efforts
together. Figure 1.7 shows an overview of the
system architecture design development
Building a GIS:
methodology. Enterprise architects now have Implementation strategy
tools they can use to complete an enterprise
wide business needs assessment.
There are four architecture domains that are
commonly accepted as subsets of an overall
enterprise business needs assessment.
These include generally accepted guidelines
and best practices provided by The Open
Group global consortium.
• The Business Architecture defines the
business strategy, governance,
Figure 1.8 System design implementation strategy
organization, and key business
moves through several steps that require planning
processes. and implementation management.
• The Information Systems Architecture
includes a review of the Data and There are several critical deployment stages
Application architecture. that support a successful implementation.
> The Data Architecture describes the Understanding the importance of each stage
structure of an organization’s logical and the key objectives for success leads to
and physical data assets and data more effective enterprise implementations.
management resources. Figure 1.8 shows a series of typical system
deployment stages for building and

maintaining successful enterprise GIS The CPT tasks include a review of business
operations. needs, establish performance targets, identify
user locations, review network suitability,
Requirements phase
select the product architecture, select the
• User information product needs establish
optimum software configuration, and complete
a foundation for completing the design.
the system architecture design analysis.
• User location and peak business loads
Additional tools are provided to validate
establish a foundation for system
system performance during the design,
architecture design.
construction, and implementation phases.
Design phase CPT models can be easily updated and
• Infrastructure requirements must be maintained to reflect changes in business
identified to quantify deployment costs. requirements and review alternative system
• Network communication capacity is an deployment strategies, providing an adaptive
important consideration for GIS model for addressing a variety of incremental
deployments. planning activities.
• Hardware and software procurement Best Practice: Build and maintain a simple
requirements must be identified. system performance model that links user
• Software development and data requirements with system design.
acquisition needs must be identified.
Best Practice: Business decisions for

System design strategies overview
project funding and procurement
authorization are often required for project
effort to proceed beyond this phase.
Construction phase
• System procurement authorization, based
on the design budget and deployment
timeline.
• Data acquisition and database design
efforts begin.
• Procurement authorization for application
design and development.
• Prototype testing plans completed and
scheduled to validate product delivery
within design performance targets.
Implementation phase
• Initial deployment and operational testing.
• Final system delivery, user training, and Figure 1.9 System design strategies overview
workflow migration complete. showing the principal chapters that will be covered
in the SDSwiki documentation and how they
• System maintenance operations. interface in building an enterprise design solution.
Enterprise GIS design includes a broad range
Best Practice: Deployment process is
of technology that must play together to satisfy
repeated incrementally on a periodic
identified business needs. The better they
schedule to leverage technology change.
work together, the more productive your
The Capacity Planning Tools (CPT) were business can be.
developed as a framework to promote
System design topics will be discussed
successful GIS system design and
throughout the SDSwiki documentation.
implementation. CPT functions contribute
throughout the implementation cycle. • Software technology: Chapter 2
• Software performance: Chapters 3 and 4

• Data administration: Chapter 5 There are three common types of user
• Network communications: Chapter 6 productivity: light, casual, and power.
• Platform performance: Chapter 7
Best Practice: Select the appropriate
• Information security: Chapter 8 productivity for each user workflow.
• Platform architecture: Chapter 9
• Performance Management: Chapter 10
• City of Rome case study: Chapter 11
General guidelines:
• System Implementation: Chapter 12
• 10 DPM is maximum user productivity for
All of the SDSwiki system architecture design
a power desktop user.
content (Chapters 1-10) are brought together
• 6 DPM is maximum user productivity for a
in a system architecture design case study for
web client workflow.
the fictional City of Rome.
• 3-5 DPM is appropriate for casual users.
Several key lessons learned are identified in • 1 DPM is appropriate for light users.
Chapter 12: System Implementation sharing
important project management and system Service time is determined by software
delivery practices that reduce implementation functions, data source, and display complexity
risk. The purpose of the system design • Light complexity is half the processing
strategies documentation is to share time of medium complexity.
fundamental practices that promote GIS • Medium complexity is the workflow
implementation success. processing baseline.
• Heavy complexity is 50 percent more than
medium complexity.
Capacity planning • Very complex workflows are possible (up
to 10x medium).
terminology
There is a great deal of terminology used Warning: Work transaction service time
within the System Design Strategies wiki that directly affects system loads.
might be new to many readers. System
Best Practice: A heavy workflow
architecture design crosses a variety of
complexity is a conservative initial
disciplines, including business operations, IT
planning performance target for most
operations, and software development. The
software technology selections.
terminology introduced in this SDSwiki works
to build a common vocabulary for
communications that bring these groups
Workflow performance
together—establishing a framework for
discussing and addressing system
performance problems.
User workflow terminology
Figure 1.11 Response time includes display

service time plus queue time.
Response time is the sum of service time and
queue time. Response time is the total time
Figure 1.10 GIS user workflow service time to complete an average display
determines display complexity.
refresh. Queue time is the average process
User workflow complexity is defined by wait time due to service contention.
service time. The user is the person working
System queue time is predictable for large
at the computer display. Service time is the
random populations.
average display refresh processing time.

• Large population: System throughput • Throughput is a measure of the
rates normally exceed several thousand transaction rate.
transactions per hour, and each • Capacity is the maximum rate at which a
transaction can include hundreds of platform can do work.
program instructions. • Utilization is the percentage of capacity
• Random distribution: System throughput represented by a given throughput rate.
loads are caused by hundreds of
Display transaction is the processing load to
transactions generated by each user.
render a new user display.
• Large random distribution of display
transactions follows a predictable random • The software program provides a set of
arrival distribution. instructions executed by the computer to
complete a work transaction.
Queue time variables include system
• The processor core executes the
utilization and transaction service time.
instructions defined in the computer
• Queue time increases with increasing program to complete the work transaction.
platform utilization.
Transactions with more instructions represent
• Queue time increases with increasing
more work, while transactions with fewer
transaction service times.
instructions represent less work.
Response time is important because it can
Throughput is average work transactions
affect user productivity.
completed over time.
• As queue time increases, response time
• Expressed in displays per minute or
will increase and user productivity may
transactions per hour.
decrease.
• An average cumulative processing load is
• Power users can be very sensitive to
applied to the server.
minor changes in display response times.
• User productivity can have a direct effect As throughput increases, the processing load
on business operations and the system's on the server increases.
ability to meet user needs.
Platform capacity is 100 percent throughput.
• Expressed in displays per minute or
Best Practice: The faster a user can work,
the more work the user can do. transactions per hour.
• Maximum server throughput is less than
platform capacity.
System performance terminology Platform utilization is the percentage of time
that the processor is busy.
• Processor is busy when servicing a
transaction.
• Processor is not busy when waiting for a
transaction.
Utilization of 60 percent means that the
processor is busy 60 percent of the time.
Figure 1.12 System performance measurements Platform utilization is the percentage of activity
show throughput and CPU utilization for a given over a period of time.
processing load
• Short sampling periods (1 second) are
System performance terminology can be difficult to evaluate.
represented by simple relationships. • Longer sampling periods (30 seconds)
• A workflow transaction (display) is an show average utilization.
average unit of work. • Long sampling periods (30 minutes) may
underestimate peak values.

• CPT includes a RESET ADJUST function
Best Practice: Time period should that will automatically reduce workflow
represent a statistically significant sample productivity to the proper reduced value.
of random transaction arrivals (100 or
more transactions). Batch process
Valid workflow Figure 1.14 Batch process productivity depends on

computed response time.
What is a batch process?
A batch process is a workflow that does not
Figure 1.13 User productivity depends on adequate
require user interaction. Most heavy GIS
system response time.
analysis can be modeled as a batch process.
What is a valid workflow?
There are several advantages of configuring
All user workflow performance terms work geoprocessing functions as a network service:
together during each display transaction to
• A service work request is sent to a
satisfy business performance requirements.
processing queue to await execution.
Workflow specifications: • A limited number of server cores can be
• User productivity = 10 DPM/client (user allocated to execute the service requests.
workflow performance needs) • Users can do other work while waiting for
• Display cycle time = 6 seconds (60 service response.
seconds in a minute divided by 10) Batch process loads are modeled for
For a given display executed on a given workflows with zero (0) think time.
platform: • Batch productivity is calculated based on
• Display service time is a constant value. computed response time (60
seconds/response time = batch DPM).
• In a shared server environment, queue
• Batch process queue time is caused by
time increases with increasing user loads
(increasing server utilization). service contention (no random arrival
queue time).
• As queue time increases, display
• Transactions are processed sequentially.
response time increases.
• Batch processes deployed on a single
• For a fixed user productivity (10 displays
per minute), computed user think time will platform with local data source tend to
decrease with increasing display consume a single processor core.
response time. • The CPT Design tab will distribute loads
across available core resources based on
For a valid workflow, the computed user think batch workflow profile.
time is greater than the minimum think time.
Warning: At some point, the computed Best Practice: Any heavy geoprocessing
user think time will be less than the function that may be requested by more
minimum think time (invalid user than one user at a time should be
workflow). separated from the user application
workflows and executed as a separate
During peak system loads, the queue time can
network batch process.
increase to a point where computed think time
is less than minimum think time. Warning: The CPT Design productivity
ADJUST function must be used to
• Workflow productivity must be reduced to compute system loads and batch process
identify a valid workflow. productivity.

GIS business planning System architecture design process
The enterprise GIS system design process
aligns identified business requirements (user
needs) derived from business strategy, goals,
and drivers (business processes) with
recommended business information systems
infrastructure technology (network and
platform) recommendations.
1. User needs assessment (results of a
GIS user needs assessment provides
Figure 1.15 System architecture design should be inputs for the system architecture design
an integral part of GIS business planning. analysis)
2. Workflow loads analysis (translate user
Figure 1.15 shows the process for GIS
needs to project workflows with baseline
planning. GIS planning provides a foundation
traffic and processing transaction loads
for building and maintaining successful GIS
based on estimated workflow complexity)
operations.
3. Technical architecture strategy (identify
Planning begins by thinking about what you user locations, network connectivity, and
want out of the system. data center server locations)
4. User requirements analysis (translate
• What is the overall business strategy?
peak user workflow loads to peak
• What are the critical business processes?
throughput transaction loads)
• What information products drive business
5. Network suitability analysis (translate
decisions?
peak site/network throughput loads to
• What are the critical data resources? peak site/network traffic and compare with
• How does GIS support current business available network bandwidth)
operations? 6. Platform architecture selection (Identify
• What are the current business needs? data center platform tier configuration and
• What are the GIS software technology identify platform selection for each tier)
needs? 7. Software configuration (Identify platform
• What are the available software assignment for each workflow software
technology deployment patterns? component peak transaction processing
• What is the process for migrating load)
business operations to the new 8. Enterprise design solution (combine all
technology state? peak workflow software component
• What are the costs and benefits for processing loads on the assigned platform
making this business change? tier, translate baseline processing load to
These are some of the questions that are selected platform processing load, and
addressed during a typical business needs generate number of nodes required for
assessment. each platform tier with estimate of
capacity utilization)
System architecture design is a process for
identifying the required network and platform The Esri Capacity Planning Tool (CPT) was
infrastructure. Network and platform developed to simplify business requirements
infrastructure must satisfy peak system data collection and automate system design
capacity and performance needs to enable analysis tasks associated with each phase of
planned business operations. the development cycle. This section will
introduce the available capacity planning tools
Best practice: System architecture design and their function in completing a system
should be included as an integral part of architecture design. Hyperlinks are provided
every business planning process. to associated CPT demos located in the
Capacity Planning Tool appendix.

The CPT can be used to identify your system
design requirements and model performance
and scalability of your enterprise design. The
CPT provides a framework for integrating
business, data, applications, and technical
architecture needs required to design, deploy,
and manage successful enterprise GIS
operations.
Pre-design efforts
Figure 1.17 GIS technology has evolved to support
a broad integrated range of business needs across
the organization. Each GIS technology pattern is
optimized to address specific organizational needs.
Figure 1.17 shows an overview of the ArcGIS
technology patterns. GIS enterprise vision
looks at how GIS technology can best support
your business needs. The ArcGIS Platform
Figure 1.16 Business needs establish the includes a range of technology options
foundation for any enterprise GIS design. The developed as a complete set of integrated
enterprise vision, existing business architecture, workflows and systems to satisfy a broad
and user requirements must be understood to range of business requirements.
select the best GIS solution.
GIS software deployment patterns are
Figure 1.16 shows how you can prepare for optimized to support your business needs:
your system architecture design. Business
needs must be understood before you are • Location enablement
ready to complete the system architecture • Data management
design. Business requirements analysis • Analysis
includes a review of the enterprise vision, the • Field mobility
existing business architecture, and the user • Visualization (Operational Awareness)
workflow requirements. Each of these areas • Constituent engagement
must be explored in some detail before you Most successful enterprise GIS operations
begin the design. evolve to embrace the full range of available
The GIS needs assessment begins with the GIS technology patterns to address focused
organization identifying where GIS technology business needs throughout their organization.
can improve the quality and productivity of the Best practice: Establishing a clear
business process flow. This assessment enterprise GIS vision early in
identifies GIS application and data planning can help identify an
requirements and an implementation strategy optimum roadmap for building
for supporting GIS user needs. The user effective GIS operations.
organization must be actively involved
throughout the user needs assessment. A GIS
solutions architect familiar with current GIS Review Business Needs
technology patterns and customer business
practices can help facilitate this planning
effort; the real work must be done by the
organization themselves. Figure 1.18 GIS technology has evolved to support
a broad integrated range of business needs across
the organization. Each GIS technology pattern is
Enterprise vision optimized to address specific organizational needs.

Figure 1.18 identifies the CPT tabs for design process as shown in Figure 1.19. Each
documenting findings from your pre-design cycle of the system architecture design
efforts process includes the following steps:
Business needs (user requirements) establish • Technical architecture strategy—High-
a foundation for building a proper system level overview showing user site
design. Business requirements are locations, network bandwidth
represented by user workflows. The first step connections, and central data center
in completing a system architecture design is locations. User location information is
to select the appropriate software deployment collected during the user needs analysis.
patterns (project workflows) and identify the • User requirements analysis—CPT
peak user workflow profiles loads. GIS Requirements analysis section is
Software Technology chapter provides an configured to represent the site locations,
overview of the primary GIS workflow patterns user workflows, peak loads, and network
along with best practices for making a proper bandwidth for the enterprise design
technology decision. solution.
• Network suitability analysis—CPT Design
CPT Workflow loads analysis
completes the network suitability analysis
The Project Workflows section on the CPT and identifies any communication
Workflow tab is located at the top of the bottlenecks. Network bandwidth upgrades
workflow list. Project Workflows are selected are identified to complete the network
from the Standard Workflows included on the suitability analysis.
CPT Workflow tab or custom workflows • Platform architecture selection—CPT
generated from the CPT Calculator or Test Design Platform tier is configured to
tabs - identifying your project workflows is the represent the design solution. Identify
process referred to as your user workflow platform tier nicknames, select platforms,
loads analysis. Once your software and identify platform rollover settings.
technology (project workflow) selection is • Software configuration—CPT Design
complete, your selected workflow Software Configuration module is used to
performance targets will be available at the assign workflow software to supporting
top of the CPT workflow selection list for use platform tier (software install) and make
in your system architecture design. workflow data source selection.
• Enterprise design solution—Once
System Design Process configured, the CPT Design tab
completes the system architecture design
analysis and provides the platform
solution.
Figure 1.19 System Architecture Design

process provides a logical step by step
methodology for using the CPT to complete
your System Architecture Design.
Once you have identified your project
workflows, you are ready to complete your
system design. The CPT is developed for use
based on a standard system architecture

Technical architecture strategy
User requirements analysis
Figure 1.21 The CPT Design tab includes

modules to address a variety of functions used
during the enterprise system architecture
design process.
Figure 1.20 Where users are located and how they Figure 1.21 shows the steps to complete the
are connected to central applications and data system design. System design process
sources can determine acceptable ArcGIS includes the user requirements analysis,
deployment patterns. network suitability analysis, platform
During your user needs assessment, you architecture selection, software configuration
must identify all user locations requiring (installation), and the enterprise design
access to GIS server applications and data solution.
sources. This should include everyone who The CPT was designed to complete the
might need access to the system during peak analysis. Once the CPT is properly configured
work periods. The term “user locations” and business requirements are defined, the
encompasses local users, remote users on CPT will complete the system architecture
the wide area network (WAN), and Internet design analysis and display the hardware
users (internal and public). solution.
The enterprise network infrastructure must be CPT Design tab: User requirements
able to accommodate peak workflow traffic analysis
loads.
Once you identify your business workflows,
• Know where users are located. you are ready to complete a user system
• Know network bandwidth communication loads analysis. The CPT Design tab includes
constraints. a Requirements Analysis module where you
• Identify the location of necessary data can identify user locations and peak
resources. throughput loads.
Network suitability analysis
Best practice: A simple diagram can help CPT Design tab: Network suitability
identify user locations and network analysis
connectivity constraints.
Once you configure the CPT Design tab to
reflect your user requirements and identify the
network connections, the CPT completes a
network suitability analysis. Network suitability
analysis is completed by the CPT Design tab.
The CPT analysis evaluates network
bandwidth and latency to ensure adequate
capacity to accommodate peak traffic flow
loads. You can then increase network
bandwidth as required to accommodate peak
traffic loads.

Platform architecture selection
CPT Design tab: Platform architecture
Monitor performance
selection compliance
Platform selection is a critical decision in any
design process. The selected platform directly
contributes to user display performance,
platform capacity, and software licensing cost.
A faster processor core improves user
performance and reduces license cost. Higher
capacity servers reduce the total number of
server machines required to satisfy business
requirements.
Warning: Platform selection directly
impacts overall system cost. Figure 1.22 Monitoring performance compliance
throughout development (map publishing), initial
Software configuration
deployment, and in production can reduce
CPT Design tab: Software configuration implementation risk.
Note: CPT Design software configuration will
be discussed in more detail in Lesson 7: GIS The capacity planning tool (CPT) can be used
product architecture. by project managers to establish specific
project performance milestones and measure
Enterprise system design compliance. Figure 1.22 shows some key
CPT Design tab: Enterprise system design opportunities for measuring performance
Once you make your platform selections and compliance.
install workflow software components • Measuring display rendering time when
(software configuration), the CPT will authoring a service is the first opportunity
complete the platform sizing and show the to validate performance.
required server configuration (number of • Measuring deployed service rendering
platforms and peak utilization levels for each time is another opportunity to validate
platform tier). performance.
The Design tab provides a snapshot view of
your user requirements and associated Best Practice: Monitoring progress in
platform solution as a single integrated meeting performance milestones can
information product. reduce deployment risk and ensure
Note: Chapter 11 provides a case study for project delivery success.
the City of Rome, demonstrating how to use When performance issues are identified early
the CPT to complete an enterprise system in deployment, proper adjustments can be
architecture design. made before impacting production workflow
CPT Hardware tab: Platform performance productivity. Simpler map displays, simpler
data models, and cached data layers are
The CPT Hardware tab is used to list SPEC proven ways to reduce workflow complexity.
performance benchmark values for use in Warnings identified by the map publishing
completing the system architecture design. analysis tool identify how to reduce display
complexity.
Best Practice: Identifying and resolving
performance issues before they become
production level performance problems
will promote deployment success.

o Platform SPECint_rate2006
baseline = 400, medium
complexity, throughput = 116,250
Figure 1.23 The CPT includes additional tools for TPH
performance validation and platform sizing. o Platform SPECint_rate2006
baseline = 100, medium
Figure 1.23 shows some tools that were complexity, throughput = 29,063
developed to translate system performance TPH
measurements to workflow service times for o Platform SPECint_rate2006
use in performance validation. The CPT Test baseline = 200, light complexity
tab was designed to help with the validation (0.5x medium), throughput =
process, and includes several tools that can 116,250 TPH
be used to translate performance o Platform SPECint_rate2006
measurements to equivalent workflow service baseline = 200, heavy complexity
times. (1.5x medium), throughput =
Performance validation (CPT Test tab) 38,750 TPH
Measuring performance during the system This section ends by introducing a simple tool
build and deployment reduces implementation used for proper hardware platform selection.
risk. System design is only the beginning of The Platform Capacity Calculator identifies the
the process; then you need to build a system single server throughput capacity of the
that performs within the system architecture selection platform configuration.
design performance targets. If measured Platform Capacity Calculator
performance loads exceed planned budgets,
adjustments can be made to workflow The Platform Capacity Calculator provides a
complexity during the build process to deliver peak throughput range for selected workflows,
services within initial performance targets. showing medium complexity output in blue
and light complexity output in red. This
provides a simple tool to help customers
Platform capacity calculations select the appropriate platform configuration.
Platform capacity for a standard map display CPT Calculator tab provides a much more
can be calculated for any selected platform. accurate platform capacity planning
Platform capacity varies based on selected assessment for a single workflow
platform throughput performance and configuration, taking into account software
published map display complexity. Vendor performance variables and hardware
published throughput benchmarks configuration strategies.
(SPECint_rate2006 Baseline) available on the
SPEC Web site shows relative platform CPT Design tab provides a much more
throughput for available server platform adaptive enterprise capacity planning
configurations. assessment integrating system loads for
multiple workflows supporting multiple network
• 2016 Baseline System Specifications local and remote sites and a variety of
o Arc16 4-core platform platform configuration strategies.
performance SPEC baseline = 256
(SPECint_rate2006 Baseline)
o ArcGIS for Server REST 2D V Concluding remarks
100%Dyn 1366x768 resolution Proper GIS planning is the most important
PNG24 output investment any organization can make in
o Baseline 4-core medium building a GIS. Understanding your GIS
complexity throughput capacity = needs, selecting the right technology at the
74,400 TPH right time, and establishing documented
• Variations in platform performance and implementation milestones to measure your
workflow display complexity drives progress can ensure your success. This
selected platform throughput document is focused on sharing how to build

and maintain successful GIS operations. The CPT Capacity Planning videos
Capacity Planning Tools provide a framework
for collecting what you know about your The next Chapter will review the most
business needs and your system common GIS technology patterns and share
environment. The Capacity Planning Tool best practices in making the right technology
models connect what you understand about selection.
GIS user requirements with the network and
platform loads your IT support teams can
measure in the data center.

GIS Software Technology
Fall 2017 GIS Software Technology 41st Edition
Since the early 1970s, Esri has continued to technology forward at an increasingly rapid
develop GIS software technology supporting pace.
functional requirements identified by the GIS
Software development history gives us
user community. Sensitivity to software
insight into the basic principles that guide us
development trends and enterprise
in building the technology of the future.
architecture strategies provide guidelines for
development investment. Esri software Technology drives software
developers leverage the latest computer innovation
hardware and software technology to
maintain leadership in the GIS marketplace. Advances in technology continues to expand
Resources are aligned to provide the best GIS capabilities.
software and services based on GIS Tightly scripted software code
customer needs.
The early ARC/INFO software provided
A variety of ArcGIS software technology developers and professional GIS users with
patterns are available to satisfy enterprise a rich toolkit for geospatial query and
business operational needs. Technology analysis and demonstrated the value of GIS
patterns include a broad range of Desktop, technology.
Server, and Mobile deployment options.
Selecting the right technology is a critical • ArcView introduced easy-to-use
step in the enterprise system design commercial off-the-shelf (COTS)
process. software that could be used directly by
GIS operational users.
GIS Software Evolution • Map Objects empowered developers with
a simple way to integrate GIS in focused
business application environments.
• Terminal servers enabled remote user
access to centrally managed GIS
desktop applications.
• ArcIMS web services introduced a
framework for publishing GIS information
products to web browser clients.
• ArcStorm and ArcSDE introduced better
ways to maintain and share GIS data
Figure 2.1 Evolution highlights of ArcGIS software from resources.
1982 to present.
Object-relational software
There is much we can learn from our past,
and Figure 2.1 provides a high-level Hardware performance improvements led to
overview of the major GIS technology more efficient programming techniques
changes over the past 25 years. Each deployed in the late 1990s.
technology advance was a tradeoff between • ArcGIS Desktop software gave users a
heavier processing loads and deploying simple and powerful application interface
software that was easier to build and for many standard GIS operations.
maintain. Faster hardware processors and • ArcGIS Server and ArcGIS Engine
improved network bandwidth provide provided developers with rich processing
opportunities for more software innovation. tools and full GIS functionality for custom
As platform and network capabilities
improve, new advances in software move
GIS Software Technology 2.1

application development and times and expand use access from a
deployment. variety of client devices.
• Distributed geodatabase management • Configurable apps enable business users
tools and replication services provided to design and deploy custom services.
better integration and sharing of • Configurable apps introduced easy-to-
geospatial data. use commercial off-the-shelf (COTS)
software that enable custom web service
Service-oriented architecture
publishing by GIS operational users.
Web technology introduced more ways to • Real time streaming updates is becoming
share data and services, introducing a part of the GIS display.
services-oriented component architecture
Distributed operations
along with interoperability standards that
enable open and adaptive applications ArcGIS Enterprise, Web GIS, and Enterprise
developed from multi-vendor component solutions provide support for hybrid
architecture. distributed operations.
• Google and Microsoft introduced pre- • ArcGIS Enterprise with on-premise Web
processed (cached) online global GIS.
basemap imagery, providing free access • ArcGIS Online organization for public
to geographic information products from collaboration.
home and mobile devices. • Infrastructure and a Service (IaaS) cloud
• Online data and services become an services for expanded data center
important extension of the GIS user capacity.
experience. • ArcGIS Industry Solutions for rapid
• Rich internet client technology improves system deployment.
display performance and server capacity.
Best Practice: Software technology
Cloud computing platform architecture migration from scripts to objects to
services and to the cloud accelerated the
Hardware virtualization, data center
rate of technology change, while
automation, and self-service cloud
increasing demands on hardware
computing provide simpler ways to
performance and network connectivity.
administer and support GIS applications and
The change in technology impacted
services.
business processes in an evolutionary
• ArcGIS online provides a cloud-based way opening new opportunities for GIS to
self-service framework for sharing support enterprise and community
intelligent maps and building an online operations, helping customers better
community basemap. understand their world, and empowering
• Portal for ArcGIS provides on-premise business with more informed decisions.
support for collaboration and sharing.
• Free web mapping tools encourage
collaboration and sharing within groups
and communities throughout the web.
• Imagery is fully integrated into ArcGIS,
including a rich set of imagery
management and analysis tools.
• ArcGIS is available for mobile phones,
tablets, and a variety of new mobile
clients.
• Commercial off the shelf (COTS)
applications reduce service deployment

shapefiles, file-based geodatabases,
Emerging technology trends personal geodatabases, SQL Express,
etc.) and accessed locally.
2. Centralized GIS. GIS data is developed
and maintained in a centralized
geodatabase shared by all departments
throughout the Enterprise. Enterprise
application workflows are established
with direct user access to dynamic
geographic information products used in
their work. GIS data resides in
centralized enterprise geodatabases
hosted by relational database
management systems (RDBMS).
Figure 2.2 GIS architecture patterns have evolved over
time with improvements in both connectivity and 3. GIS online sharing. ArcGIS Online
infrastructure technology. shares a selection of common Cloud-
Figure 2.2 shows a four-quadrant chart hosted cached basemaps and imagery
sharing the relationship between technology that GIS users can access across a wide
and advancing GIS capabilities. GIS is area network (WAN). Communities
evolving as a platform for understanding our contribute original basemap resources to
world. Geography helps us visualize our data enhance the ArcGIS Online cached data
in a unique way. GIS analysis identifies sources. Data packages, model
relationships between objects in space and packages, and intelligent maps can be
time for intelligent spatial thinking. GIS shared through a secure networking
analysis requires access to large volumes of portal connecting groups and
data that have location or place information organizations promoting collaboration
tags: i.e. addresses, place names, or lat/long within and among geospatial
attributes. The result of geospatial analysis is communities. Increasing Internet
presented in a language we can all easily connectivity is a catalyst for bringing
understand: Maps. users and organizations together to solve
community geospatial problems by
GIS architecture patterns include (1) sharing better ways to manage our world.
traditional GIS departments working with GIS data is stored in the online cloud and
local data on local desktop workstations to accessed as either static map services
(4) ubiquitous collaborative environments (basemaps cached content) or as GIS
with direct access to services and online files. These files can be shared and
data sources accessed from multiple devices downloaded to be viewed or analyzed
from any location through the Cloud. using local GIS desktop software.
1. Traditional GIS. Professionals in GIS 4. Web GIS Architecture. ArcGIS Online
departments develop and maintain data Organization services extended GIS user
sources and create geographic access to directly create and share
information products from their local work geospatial data services to users on any
environment. GIS users request maps device at any location through the Cloud.
prepared by a GIS analysis for use in Direct access to dynamic geospatial data
their work. Departments throughout the services, configurable applications for
Enterprise establish their own local GIS Web and mobile phone clients, and real
groups for building and managing time geospatial data feeds over the
operational level geographic content. GIS WAN. Dynamic business layers
data is distributed as files (such as maintained by each organization can

overlay (mashup) on shared basemaps • Create: Creating and using maps.
delivering rich high performance • Manage: Managing geographic
operational geographic information information in a database.
products. Web services are hosted on a • Visualize: Using maps and geographic
self-service scalable infrastructure that information in a range of applications.
establishes ArcGIS as a platform for • Analyze: Analyzing geospatial
geospatial information collaboration and information.
sharing. GIS information products are • Collaborate: Compiling geographic data
published as dynamic, REST-based, into real time information products.
loosely-coupled data services that can be
You can work with ArcGIS using a range of
consumed by all GIS applications:
clients.
Desktop, Web or Mobile.
• Web: Rich internet application clients
ArcGIS Online Organizations enable
connecting to web-based services.
ubiquitous access and integrates the
• Mobile: Mobile devices connected over
traditional work of geospatial professionals
wireless broadband connections.
with a whole new world of GIS applications.
• Desktop: Workstations connected to data
It takes what have been relatively scarce
sources throughout the enterprise.
commodities—stories and actionable
geoinformation—and makes them abundant. These services can be hosted in a range of
Web maps provide the medium for settings:
integration and understanding and make this • Cloud services: Published and shared in
information widely accessible in simple the cloud.
forms. This widespread, easy access to • Enterprise services: Web services for use
geographic knowledge is what we mean by within an enterprise.
providing geography as a platform for • Local services: On a local computer (and
understanding. as files on disk).
Best Practice: Increased connectivity and ArcGIS provides an infrastructure for making
integration promote use of dynamic GIS maps and geographic information available
content. throughout an organization, across a
What is ArcGIS? community, and openly on the web. Each
GIS service can be shared within a specific
workgroup or organization; within a small,
well-defined user community; or publicly on
the open web.
Standard ArcGIS deployment patterns:
• Departmental: Local work area desktops
connected to local data servers.
• Centralized: Centralized geodatabase
Figure 2.3 What is ArcGIS today? server accessed by multiple departments
located throughout the enterprise.
Figure 2.3 shares how GIS is getting easier • Enterprise: Enterprise operations
to use, more powerful, and available in more distributed over local, remote, and cloud
places today than ever before. architecture.
ArcGIS shares a global platform for working • Federated: Distributed GIS operations
with maps and geographic information. It is sharing spatial data sources via web
used to… portals or replicated geodatabase
services.
• Discover: Sharing and discovering
geographic information.

• Service-oriented architecture: Workflow • ArcGIS Desktop clients create and
integration through web services and publish services content.
messaging. • GIS content (web layers) can be
• Cloud computing: Hosted self-service published and shared from server, online
data center operations. content, and Internet services.
• Web maps can be developed by end
GIS tends to bring business units and
users for analysis and collaboration.
organizations together to solve common
• Access and identify management can be
problems, and as such is established as an
managed by the end user community.
important component of enterprise
• GIS information products (web maps)
operations management.
can be accessed as a service and shared
Best Practice: Enterprise system design on any client application.
is about selecting the right software
deployed on the best platform providing ArcGIS platform use role
the best return on investment.
summary
ArcGIS Enterprise
Figure 2.5 Standard use cases supported by GIS

operations.
Figure 2.5 shows the standard user roles

supported by the ArcGIS platform.
GIS business needs can be represented by a
variety of use cases represented by the
following ArcGIS platform user roles.
ArcGIS Desktop use cases
• Expert: GIS professionals and data
Figure 2.4 ArcGIS Enterprise enables content scientists who perform advanced
development and management at a business user geospatial analysis, visualization,
level.
mapping, and data management.
Figure 2.4 shows the ArcGIS Enterprise • Curator: GIS professionals who curate
deployment concept, managing information and maintain authoritative maps and
from a community of services and delivering geospatial databases.
web maps to applications supported on • Analyst: People who perform analysis,
every device. visualization, and mapping.
ArcGIS portal (ArcGIS Online and Portal for Web apps use cases
ArcGIS) content management enables • Contributor: Knowledge workers who
customer engagement, collaboration, and
create, update, and share maps and
sharing, which empowers business users to related content.
create and share content (information • Viewer: Everybody who views and
products). interacts with existing maps.
Integrated platform architecture facilitates Best Practice: Standard ArcGIS licensing
information development and sharing. options are provided for each user role.

ArcGIS Curator
ArcGIS Desktop user roles
ArcGIS Desktop user roles include ArcGIS
Expert, ArcGIS Curator, and ArcGIS Analyst.
ArcGIS Expert
Figure 2.7 ArcGIS Curator role.
Figure 2.7 shows the ArcGIS Curator user

role. The ArcGIS Curator user role includes
the following capabilities:
Software capabilities
• ArcGIS Desktop Standard
Figure 2.6 ArcGIS Expert role. • Apps for Field and Office
• Basic App Builders and Solution
Figure 2.6 shows the ArcGIS Expert user Templates
role. The ArcGIS Expert user role includes • L2 Read-Write Identify (when paired with
the following capabilities: ArcGIS Online or Enterprise)
Software capabilities Licensing
• ArcGIS Desktop Advanced • Perpetual and term licensing options
• Apps for Field and Office
• Basic App Builders and Solution Optional extensions
Templates • Advanced Analysis: ArcGIS 3D Analyst,
• L2 Read-Write Identify (when paired with ArcGIS Geostatistical Analyst, ArcGIS
ArcGIS Online or Enterprise) Network Analyst, ArcGIS Schematics,
Licensing ArcGIS Spatial Analyst, ArcGIS Tracking
Analyst
• Perpetual and term licensing options. • Data and Workflows: ArcGIS Data
Optional extensions Interoperability, ArcGIS Data Reviewer,
ArcGIS Publisher, ArcGIS Workflow
• Advanced Analysis: ArcGIS 3D Analyst, Manager
ArcGIS Geostatistical Analyst, ArcGIS
Network Analyst, ArcGIS Schematics,
ArcGIS Spatial Analyst, ArcGIS Tracking ArcGIS Analyst
Analyst
• Data and Workflows: ArcGIS Data
Interoperability, ArcGIS Data Reviewer,
Manager
Figure 2.8 ArcGIS Analyst role.

Figure 2.8 shows the ArcGIS Analyst user Midterm (Q4 2017)
role. The ArcGIS Analyst user role includes
• Utility Network, Attribute Rules, Reports
• Offline Web Maps & Layers, Stereo &
Software capabilities Oblique
• Exploratory Spatial Data Analysis
• ArcGIS Desktop Basic
• Real-Time Streaming, SAP HANA
• Basic App Builders and Solution Long-term (2018)
Templates
• Parcel Fabric, Offset Printing,
• L2 Read-Write Identify (when paired with
Presentations
ArcGIS Online or Enterprise)
• Animated Symbols, Dimensions
Licensing
ArcGIS Pro will soon be feature equivalent
• Perpetual and term licensing options with ArcMap, with more advanced
capabilities and professional GIS user
Optional extensions
productivity tools.
• Advanced Analysis: ArcGIS 3D Analyst,
Best Practice: Plan now for leveraging
ArcGIS Geostatistical Analyst, ArcGIS
ArcGIS Pro capabilities
Network Analyst, ArcGIS Schematics,
ArcGIS Spatial Analyst, ArcGIS Tracking
Analyst
• Data and Workflows: ArcGIS Data
3D visualization
Interoperability, ArcGIS Data Reviewer,
Manager
ArcGIS Desktop is changing
Figure 2.10 3D visualization is available in CityEngine,

ArcGIS Pro, and ArcScene.
Figure 2.9 ArcGIS Pro is expanding ArcGIS Desktop Figure 2.10 shows advances with 3D
capabilities. visualization. 3D models, animations,
Figure 2.9 shows the rapidly advancing renderings, virtual tours, and planning are
ArcGIS Pro road ahead. some of the services available with the
ArcGIS platform. ArcGIS Pro provides
Near term (July 2017) powerful 3D capabilities for development and
• Multiple Instances of Pro design, with options for 2D and 3D
• Annotation Editing, Read WFS, Layout geographic representations in an interactive
Grids linked multi-display view.
• Full Motion Video, Charts in Layouts Create interactive scenes with the built-in
• Interactive Input for Analysis, SDK Scene Viewer. A scene is symbolized
Improvements geospatial content that includes a multiscale

basemap and a collection of 2D and 3D Vector tiles are created and delivered with
layers, styles, and configurations that allow much less processing.
you to visualize and analyze geographic
• Faster tile generation: World vector tiles
information in an intuitive and interactive 3D
generated in 8 hours versus world raster
environment.
tiles generated in many weeks.
1. Choose global or local • Reduced tile size: World vector tiles ~13
2. Switch basemap GB; world raster tiles ~20 TB.
3. Add layers
Vector tiles are published and consumed by
4. Configure layers
ArcGIS client software.
5. Capture slides
6. Save scene • Tile creation: ArcGIS Pro 1.2+
(geoprocessing services).
Best Practice: 3D visualization can bring
• Published tile layers: ArcGIS Online;
your ideas to life.
ArcGIS Server/Portal for ArcGIS 10.4+.
New vector tiles • Client consumption: ArcGIS Runtime
Quartz; ArcGIS Pro 1.3+, ArcGIS
JavaScript 3.15+ and 4.0 APIs.
Best Practice: Vector tiles provide a
single layer of files that can be
dynamically displayed and styled.
ArcGIS Desktop migration
Figure 2.11 ArcGIS is evolving to consume vector tiles

across all devices as well as create and publish vector
tiles to ArcGIS Server and ArcGIS Online.
Figure 2.11 shows the new vector tile format.

Esri is engineering support for consuming
and rendering vector tiles across the ArcGIS
Figure 2.12 Enterprise ArcGIS Desktop solutions
platform. ArcGIS Pro can create, publish, should consider migration to ArcGIS Pro.
and use vector tile basemaps for optimum
ArcGIS Desktop and Web GIS performance Figure 2.12 shows ArcGIS Desktop migration
and scalability. strategy. ArcGIS Desktop users are
migrating from ArcMap to ArcGIS Pro.
Vector tiles support a more dynamic high-
quality display. Legacy applications
• Display quality: Best possible resolution • Most system of record environments
for Retina displays; small, efficient currently use ArcMap for their production
format. environments.
• Dynamic labeling: Clearer, more • ArcMap remote application server load
readable text; on-the-fly labeling for profiles are different than ArcGIS Pro.
heads-up display. ArcGIS Desktop future
• Map styling: Streets, Topo, Canvas from
one tile set; Day and Night mode; • ArcGIS Pro should be able to replace all
Restyling. ArcMap capabilities by2019.
• ArcGIS Desktop license supports both
ArcMap and ArcGIS Pro.

• ArcGIS Pro remote application server • ArcGIS Desktop is installed on the client
configurations (Citrix) are different than workstation with direct data source
ArcMap. access.
• The ArcGIS Desktop application
Best Practice: System architecture
connects to a file or Enterprise
should consider migration to ArcGIS Pro
Geodatabase data source.
in their final design solution.
AGD wkstn FSvc workflows
Note: ArcMap and ArcGIS Pro configuration
guidelines will be discussed in chapters 3 • ArcGIS Desktop is installed on the client
and 9. workstation with access to a GIS Server
feature service.
ArcGIS Desktop workflows • The ArcGIS Desktop application
ArcGIS Desktop business use cases are connects to a GIS Server feature service.
represented in the Capacity Planning Tool AGD_wkstn Imagery workflows
(CPT) as ArcGIS Desktop workflows. ArcGIS
• ArcGIS Desktop installed on client
wkstn workflows represent ArcGIS Desktop
workstation accessing an imagery data
software installed on the client workstation.
source.
ArcGIS Citrix workflows represent ArcGIS
• The ArcGIS Desktop application
Desktop software installed on a central host
connects to an imagery data source.
server farm (windows terminal servers)
supporting distributed thin client access and Baseline ArcGIS Desktop workstation
control. workflows are included in the Capacity
Planning Tool.
ArcGIS Desktop use case
• Standard workflows are included in the
The following use cases support ArcGIS
CPT Workflow tab.
Desktop workflows.
• Custom workflows can be generated
ArcGIS Desktop wkstn workflows using the CPT Calculator tab.
• Project workflows are included in the
CPT Workflow tab.
Best Practice: Workflow performance
targets are generated by the CPT
Calculator from benchmark test results.
ArcGIS Desktop Citrix workflows
Figure 2.13 AGD wkstn, AGD wkstn FSvc, and

AGD_wkstn Imagery workflows are included in the
CPT for capacity planning.
Figure 2.13 shows the types of ArcGIS

Wkstn workflows available in the CPT.
AGD wkstn workflows
Figure 2.14 AGD Citrix, AGD Citrix FSvc, and
AGD_Citrix Imagery workflows are included in the CPT
for capacity planning.

Figure 2.14 shows the types of ArcGIS Citrix Best Practice: Workflow performance
workflows available in the CPT. AGD Citrix targets are generated by the CPT
workflows are used to represent remote Calculator from benchmark test results.
client access to centrally hosted ArcGIS
Desktop applications.
ArcGIS Desktop workflows for
capacity planning
AGD Citrix workflows
ArcGIS Desktop software technology
• ArcGIS Desktop application installed on selection
host server farm with direct data source
access. • ArcGIS Desktop workflow performance
• The terminal client connects to the targets are generated from the CPT
ArcGIS Desktop application installed on Calculator tab.
the host platform. • The CPT Calculator generates a
• The ArcGIS Desktop application on the workflow recipe that identifies the
host connects to the file or Enterprise selected software technology and
Geodatabase data source. performance parameters.
AGD Citrix FSvc workflows CPT desktop workflows
• ArcGIS Desktop installed on host server ArcGIS Desktop workflow performance

with access to a GIS Server feature targets are generated from the CPT
service. Calculator tab.
• The terminal client connects to the Software workflow recipe
ArcGIS Desktop application installed on
the host platform. The CPT Calculator generates a workflow
• The ArcGIS Desktop application on the recipe that identifies the selected Software
host connects to the GIS Server feature Technology and Performance Parameters.
service. Note: Workflow performance targets will be
AGD_Citrix Imagery workflows discussed in greater detail in Chapter 3: GIS
software performance.
• ArcGIS Desktop installed on host
platform accessing an imagery data ArcGIS Desktop Standard Workflows
source. The most common ArcGIS Desktop workflow
• The terminal client connects to the patterns are pre-generated from the
ArcGIS Desktop application installed on Calculator tab and included on the Workflow
the host platform. tab as Standard Workflows.
• The ArcGIS Desktop application on the
host connects to the imagery data
source. ArcGIS Desktop performance
Baseline ArcGIS Desktop workstation The ArcGIS Desktop workflow performance
workflows are included in the Capacity will vary, based on complexity and data
Planning Tool. source.
• Standard workflows are included in the • Workflow display complexity and user
CPT Workflow tab. productivity are two key variables that
• Custom workflows can be generated contribute to hardware platform
using the CPT Calculator tab. processing loads.
• Project workflows are included in the • Network bandwidth and display traffic are
CPT Workflow tab. two additional key variables that impact
display performance.

• The selected GIS data source format is ArcGIS Viewer
another key variable impacting display
performance.
Note: Workflow complexity will be discussed
in greater detail in Chapter 3: GIS software
performance.
Warning: The use of 3D dynamic
workflows can significantly increase
display complexity and client traffic.
Web app user roles Figure 2.16 ArcGIS Viewer role.
Figure 2.16 shows the ArcGIS Viewer user

Web app user roles include ArcGIS
role. The ArcGIS Viewer user role includes
Contributors and ArcGIS Viewers.
ArcGIS Contributor
Includes
• Read-only Apps for Field and Office
• Read-only Apps
Licensing
• Term licensing
• Volume discounts (>500,>1000)
Figure 2.15 ArcGIS Contributor role.
Figure 2.15 shows the ArcGIS Contributor ArcGIS Enterprise licensing

user role. The ArcGIS Contributor user role
includes the following capabilities:
Includes
Templates
Licensing
• Term licensing
• Volume discounts Figure 2.17 ArcGIS Enterprise provides the software
(>50,>100,>500,>1000) components to support Web GIS operations.
Figure 2.17 shows the ArcGIS Enterprise

components. ArcGIS Enterprise provides
licensing for the basic ArcGIS Enterprise
software components.
ArcGIS Enterprise licensing levels
• ArcGIS Enterprise Advanced
• ArcGIS Enterprise Standard

ArcGIS Enterprise per core licensing model Figure 2.19 The ArcGIS Enterprise Standard
license provides basic server capabilities.
• Licensing is based on the number of GIS
Extensions are available to increase
Server processor cores.
capabilities as needed to satisfy business
• Customer can choose to base licensing
requirements.
on physical or virtual server core.
• Number of GIS Server processor core Includes
limits peak system transaction
• 4-core License of GIS Server plus Portal
throughput.
• 5 – ArcGIS Contributor Role Users
• 30 – ArcGIS Viewer Role Users
ArcGIS Enterprise Advanced Licensing
• Perpetual and term licensing options
Optional extensions
• ArcGIS Schematics, ArcGIS 3D Analyst,
ArcGIS Spatial Analyst, ArcGIS
Geostatistical Analyst, ArcGIS Network
Analyst, ArcGIS GeoEvent
• Perpetual or Term (1-year) license
Figure 2.18 ArcGIS Enterprise Advanced license.
Scalability
Figure 2.18 The ArcGIS Enterprise
Advanced license provides the most server • Additional 2-core Packs
capabilities. All extensions are included in • Perpetual option with annual
the Advanced license. maintenance
Includes ArcGIS Enterprise use cases
• 4-core License of GIS Server and Portal ArcGIS web app business use cases are
• Advanced Analysis: ArcGIS Schematics, represented in the CPT as ArcGIS Server
ArcGIS 3D Analyst, ArcGIS Geostatistical (AGS) workflows. Web applications consume
Analyst, ArcGIS Network Analyst services published by ArcGIS Server. AGS
• 50 – ArcGIS Contributor Roles Users workflows include web map services, feature
• 30 – ArcGIS Viewer Role Users service, image service, and cached tile
Licensing services. Web apps can include a mix of
these services (map, feature, image, and
• Perpetual and term licensing options cached tile services) in a composite
Scalability workflow. Traditional web mapping services
were published directly from ArcGIS Server
• Additional 2-core Packs or similar web service content providers.
• Perpetual option with annual ArcGIS Enterprise includes access to Portal
maintenance content, where Web maps are created and
ArcGIS Enterprise Standard shared to define the services that are
consumed by the web app display.
Figure 2.19 ArcGIS Enterprise Standard license.

ArcGIS web apps workflows • Image Server service instances connect
with mosaic dataset to image file data
source.
Baseline workflows are included in the
Capacity Planning Tool.
• Standard workflows are included in the
CPT Workflow tab.
• Custom workflows can be generated
using the CPT Calculator tab.
• Project workflows are included in the
CPT Workflow tab.
Figure 2.20 ArcGIS Server web mapping services,

feature services, and imagery services workflows are
ArcGIS Server workflows for capacity
included in the CPT for capacity planning. planning
Figure 2.20 shows the types of ArcGIS web ArcGIS Server software technology selection
app workflows available in the CPT. The • ArcGIS Server workflow performance
following ArcGIS web app use cases are targets are generated from the CPT
included in the Capacity Planning Tool. User Calculator tab.
requirements are identified as peak • The CPT Calculator generates a
concurrent users or peak transaction workflow recipe that identifies the
throughput. selected software technology and
AGS (REST, KML, WMS, SOAP, performance parameters.
KMX)(JPEG, PNG, ..., PDF) workflows CPT server workflows
• Web Map Service published on GIS ArcGIS Server workflow performance targets
Server site. are generated from the CPT Calculator tab.
• The web client connects to GIS mapping
services published from the GIS Server ArcGIS Server standard workflows
platform. The most common ArcGIS Server workflow
• GIS Server service instances connect to patterns are pre-generated from the
the file or Enterprise Geodatabase data Calculator tab and included on the Workflow
source. tab as Standard Workflows.
AGS (REST,...) Feature workflows Best Practice: Use the CPT Calculator to
• Web feature services published on GIS complete a workflow analysis for each
Server site. business use case, and then use the
• The web client connects to GIS feature Calculator-generated workflow
services published from the GIS Server performance targets for your design
platform. specifications.
• GIS Server service instances connect to Standard workflows provide a reasonable
the file or Enterprise Geodatabase data performance target for light and medium
source. software deployment use-cases. In most
AGS Imagery workflows cases, Standard Workflows provide
performance targets with adequate margins
• Web imagery services published on for your design.
Image Server site.
• The web client connects to GIS image Best Practice: The workflow recipe
services published from the Image should be used as design specifications
Server platform. by the service author and software

developer to ensure compliance with • Data Appliance 5.0 for ArcGIS includes
performance targets during system world Base Maps, Geocoding services,
deployment. and world imagery datasets.
• Image Management Services: ArcGIS
Note: Workflow display complexity will be
Image Server provides Image Services
discussed in more detail in Lesson 3.
and Raster processing capabilities.
ArcGIS Server performance • Real Time Services: ArcGIS GeoEvent
Client traffic bandwidth limitations contribute Server provides real time capabilities.
• Big Data Analytics Services: ArcGIS
most to web service display performance.
GeoAnalytics Server provides analysis of
• Server processing loads are less than a variety of big data sources.
one second for most web mapping
workflows. GIS Server roles
• Local high bandwidth client display
performance is much faster than more
bandwidth-challenged remote client
display performance.
• Bandwidth connectivity can provide a
significant contribution to user
productivity. Figure 2.22 GIS Server roles provide distributed
processing for a variety of geoprocessing tasks within
Warning: Server processing times will the ArcGIS Enterprise architecture.
impact platform capacity and licensing Figure 2.22 shows the available GIS Server
costs, with less impact on user roles. GIS Server roles provide workflow
productivity. separation for Imagery raster analytics,
Note: The default workflows shown here use ArcGIS GeoEvent real time processing,
a 1024x768 map display resolution—display GeoAnalytics big data analysis, and
traffic and client response time improves Business Analysis demographic resources.
significantly with smaller resolution image GIS Server roles
services.
• ArcGIS Image Server: Provides
dedicated GIS Server site for image
Additional capabilities for services and raster analytics within the
ArcGIS Enterprise architecture.
ArcGIS Enterprise • ArcGIS GeoEvent Server: Provides
dedicated GIS Server site for real-time
processing within the ArcGIS Enterprise
architecture.
• ArcGIS GeoAnalytics Server: Provides
dedicated GIS Server site distributed
processing for big data analysis within
the ArcGIS Enterprise architecture.
Figure 2.21 Additional capabilities for ArcGIS • ArcGIS Business Analyst Server:
Enterprise. Provides dedicated GIS Server site for
Figure 2.21 shows the available ArcGIS business analysis within the ArcGIS
Enterprise extensions. ArcGIS Enterprise Enterprise architecture.
can be expanded to include additional Best Practice: Workload separate of
capabilities. geoprocessing tasks improve overall
Following capabilities are available with the system performance and scalability.
ArcGIS 10.5 release:

ArcGIS Image Server and raster analytics ArcGIS GeoEvent Server
Figure 2.23 ArcGIS Image Server deployment

architecture.
Figure 2.23 shows the ArcGIS Image Server

product architecture. ArcGIS Image Server
Figure 2.24 ArcGIS GeoEvent Server deployment
provides a dedicated site for publishing architecture.
imagery services and supporting raster
analytics processing. Figure 2.24 shows the ArcGIS GeoEvent
Server deployment architecture. ArcGIS
The ArcGIS Image Server role is deployed GeoEvent Server enables real-time event-
as a separate GIS Server site. To use based data streams to be integrated as data
ArcGIS Image Server, you must have the sources in your enterprise GIS.
basic ArcGIS Enterprise deployment. The
basic ArcGIS Enterprise environment Event data can be filtered, processed, and
supports Portal for ArcGIS, a Federated sent to multiple destinations, allowing you to
hosting GIS Server, and the relational and connect with virtually any type of streaming
tile cache ArcGIS Data Stores. An additional data and automatically alert personnel when
Raster data store is deployed with the specified conditions occur, all in real time.
ArcGIS Image Server site. With ArcGIS GeoEvent Server, you can
ArcGIS Image Server hosts published perform the following tasks:
dynamic image services and capabilities for • Stream (push) event data to your client
executing a wide range of raster processing applications via WebSockets.
models. The results of image processing or • Direct event data into feature services
raster analysis are persisted as image and hosted on ArcGIS Online, Portal for
feature layers registered in the portal. ArcGIS, or ArcGIS Server so that maps
that you create will represent the most
up-to-date information occurring in the
real world.
• View the latest feature status using any
ArcGIS viewer (for example, Operations
Dashboard for ArcGIS).
• Filter GeoEvents using spatial or attribute
conditions to focus on the most
interesting event data.
• Geofence areas of interest using existing
feature data to detect the spatial
proximity of events. You can even create
geofences on-the-fly without

disconnecting from your real-time data Spatiotemporal ArcGIS Data Store. ArcGIS
stream. GeoAnalytics Server machines are deployed
• Archive event data in feature services, in a separate dedicated GIS Server site.
tables, and the spatiotemporal big data
Beginning at ArcGIS 10.5, you can perform
store.
feature analysis using distributed computing
• Enrich incoming events with data from a
with the tools and platforms provided by the
secondary feature service or system file.
ArcGIS GeoAnalytics Server site. ArcGIS
To use ArcGIS GeoEvent Server, you must GeoAnalytics Server can perform data
have the basic ArcGIS Enterprise analysis from feature services, Big Data File
deployment. The base ArcGIS Enterprise Shares, or the Relational and Spatiotemporal
environment includes Portal for ArcGIS, a ArcGIS Data Stores. ArcGIS GeoAnalytics
Federated hosting GIS Server, and the Server output is provided as Web GIS layers
Spatiotemporal ArcGIS Data Store. ArcGIS in Portal.
GeoEvent Server machines are deployed as
separate GIS Server sites.
ArcGIS Business Analyst Server
ArcGIS GeoAnalytic Server
Figure 2.26 ArcGIS Business Analyst Server

Figure 2.25 ArcGIS GeoAnalytics Server deployment deployment architecture.
architecture.
Figure 2.26 shows the ArcGIS Business
Figure 2.25 shows the ArcGIS GeoAnalytics Analyst Server deployment architecture.
Server deployment architecture.
ArcGIS Business Analyst Server is part of an
ArcGIS GeoAnalytics Server introduces enterprise solution that enables
distributed computing to vector-based organizations to incorporate location into
feature data and can be used to analyze big their decision-making process.
data or accelerate traditional ArcGIS
Desktop analysis workflows through ArcGIS ArcGIS Business Analyst Server comes with
Pro and Portal for ArcGIS. several Business Analyst-based
geoprocessing tools that can be used in
The ArcGIS GeoAnalytics Server site Python-based scripts or models, and
enables distributed analysis across a published as reusable custom workflows.
scalable architecture. To use ArcGIS Additionally, the solution enables
GeoAnalytics Server, you must have the organizations to host the Esri
basic ArcGIS Enterprise deployment. The GeoEnrichment service on-premises and
base ArcGIS Enterprise environment behind their firewalls.
includes Portal for ArcGIS, a Federated
hosting GIS Server, and the relational and

ArcGIS Business Analyst Server will work Standard mapping or feature services can be
alongside single-machine and multi-tiered used to represent a geoprocessing service
deployments of ArcGIS Enterprise. The profile.
Business Analyst web app requires a portal
• Use any Standard or Custom workflows
with a hosting server; however, stand-alone
and set min think time = 0.
ArcGIS Server implementations of ArcGIS
• Identify number of batch instances as
Business Analyst Server are also supported
workflow clients.
for legacy custom applications.
• CPT calculates loads for single threaded
batch process across installed platform
profile.
ArcGIS geoprocessing
CPT Calculator batch workflows
workflows
ArcGIS geoprocessing workflows can be
very compute intensive, and can often Heavy geoanalytics processing loads
generate background platform processing
loads over a long period of time.
• Some geoprocessing tasks generate
processing loads across multiple platform
tier along with associated network traffic Figure 2.28 Dedicated host standard GeoBatch
transactions. (i.e. geodatabase Services.
replication, mobile synchronization, data Figure 2.28 shows CPT standard workflows
migration, etc). that apply batch processing loads to a single
• Many geographic geoprocessing tasks software component.
are executed on a single platform tier (i.e.
geographic spatial analysis, geographic For many geoprocessing analysis tasks, the
temporal analytics, geographic network processing is performed on a single server
tracing analysis, and imagery raster platform in memory or with a data source on
analytics). local storage.
• Geoprocessing productivity is limited by • During process execution, the
available software service instances and geoprocessing service instance
hardware processing resources. consumes a processor core.
Warning: Geoprocessing jobs consume • For capacity planning purposes, the
assigned platform resources system design architect must reserve a
platform processor core for each
Distributed geoprocessing services concurrent processing job. The
processing job output loads are negligible
compared to the batch processing load,
and can be ignored.
• Geoprocessing loads can be represented
by using a CPT batch workflow (min think
Figure 2.27 Any workflow profile can be deployed as a time = 0) with service times assigned to
batch process.
single software component (i.e. Citrix,
Figure 2.27 shows a standard ArcGIS Server Web, SOC, or DBMS).
use case configured as a geoprocessing
CPT batch Standard Workflows
service. Capacity Planning Tool
geoprocessing service loads are enabled by
setting workflow min think time = 0.

sites. Field staff can continue to work
Geoprocessing workflow productivity even when disconnected from your
network.
• Insights for ArcGIS: An app that you
access through Portal for ArcGIS that
Figure 2.29 Geoprocessing productivity.
allows you to perform iterative and
Figure 2.29 shows a representation of exploratory data analysis on ArcGIS web
geoprocessing batch process productivity. services, Excel spreadsheets, and data
stored in databases.
Geoprocessing tasks are executed as rapidly
• Drone2Map for ArcGIS: Allows you to
as possible given the available processing
visualize and analyze geolocated images
resources.
that you collect with a drone to help you
• Process runs as long as it takes to inspect areas from the air and monitor
complete the processing job. changes. 2D and 3D maps that you
• Single-threaded analysis tasks are create with Drone2Map for ArcGIS can
executed on a single process thread be shared to Portal for ArcGIS.
which can consume a single processor • ArcGIS Business Analyst web app:
core. Enables commercial Realtors, retailers,
• Multi-threaded analysis tasks can be economic development agencies, and
distributed across available software others to evaluate markets for expansion,
service instances, where each service consolidation, or investment.
instance can consume a single processor • Community Analyst: Includes global data
core. and reports for more than 135 countries.
Geoprocessing workflow performance These datasets range from population
and households to healthcare spending
Capacity planning is based on peak number and crime rates. Access recent
of concurrent batch processes required for demographics about total population,
each geoprocessing workflow (business family size, household income,
requirements). Analysis run time depends on unemployment, and more.
the complexity of the geoprocessing task and • GeoPlanner for ArcGIS: Allows you to
the performance of the selected platform collaborate with other members of your
architecture. portal organization on land-based
planning and designs.
Premium apps for ArcGIS A variety of Premium apps for ArcGIS are
available on the ArcGIS web site.
Figure 2.30 Premium apps for ArcGIS.
Figure 2.30 shows the available premium

apps for ArcGIS. Premium apps are licensed
separately from ArcGIS Enterprise. You
purchase and import entitlements for these
apps into your portal, and provision individual
licenses to portal members.
ArcGIS Enterprise premium apps
• Navigator for ArcGIS: A mobile app that
allows you to route field-workers to job

Insights for ArcGIS ArcGIS Online Services provide instant and
reliable access to terabytes of data including
street maps, live weather and traffic
information, extensive demographic data,
topographic maps, and high-resolution
imagery from an extensive list of world-class
data providers.
ArcGIS Online provides Web GIS software
as a service (SaaS).
Includes
Templates
• 5 – ArcGIS Contributor Role Users
Figure 2.31 Insights for ArcGIS. Licensing
Figure 2.31 shows the Insights for ArcGIS • Term licensing

information display interface. The Insights for • Additional Contributor Role Users in
ArcGIS app is accessed through Portal for blocks of five (5)
ArcGIS and allows you to perform iterative • Viewer Role Users in blocks of ten (10)
and exploratory data analysis on ArcGIS web • Volume discounts available
services, Excel spreadsheets, and data
stored in databases.
ArcGIS Online for Organizations
With Insights for ArcGIS, you can quickly
discover the secrets that your data holds.
Visualize, analyze, and tell your story like
never before.
Maps, charts, and tables appear alongside
each other as cards. Each card is a live
snapshot of the analysis. Apply a different
style to each card, and keep iterating.
ArcGIS Online
Figure 2.33 ArcGIS Online subscription services.

ArcGIS Online provides a variety of data content and
analysis services along with an organization portal for
managing creation, sharing, and collaboration of online
services.
Figure 2.33 shows the ArcGIS Online for

Figure 2.32 ArcGIS Online licensing.
Organizations user interface. ArcGIS Online
Figure 2.32 shows the available ArcGIS for Organizations include a GIS Portal that
Online licensing options. ArcGIS Online connects end-users with useful online GIS
services offer a cost-effective way to access applications, maps and geographic data,
up-to-date GIS content and capabilities on extending the reach of your geographic
demand. information products and ArcGIS beyond

GIS experts. Portal is a central component of GIS Software Selection
the ArcGIS platform, providing the means to
easily create, organize, secure, and manage
geographic assets within your organization.
With ArcGIS Online subscription services,
data storage, maintenance, and updates are
handled by Esri, providing software as a
service for data management, sharing,
collaborating, and publishing online. Users
can access data and GIS capabilities directly
using ArcGIS Desktop or use ArcGIS Online
subscription services to build unique Web-
based applications.
Training presentation on Best Practices for Figure 2.34 GIS enterprise solutions often include a
Your ArcGIS Online Organization mix of software technology and data sources carefully
selected to satisfy specific operational business needs.
ArcGIS Online downloadable apps Selecting the right software and the most
A number of ready-to-use apps are available effective deployment architecture is very
for download fully integrated with ArcGIS important. ArcGIS technology provides many
Online Organizations. A sample of the more alternative architecture solutions and a wide
popular apps include: variety of software, all designed to support
specific user workflow needs as shown in
• Collector for ArcGIS Figure 2.34.
• Location Analytics clients
>Esri Maps for Office GIS software technology alternatives.
>Esri Maps for Dynamics CRM • What are the best data sources?
>Esri Maps for IBM Cognos • What user workflows should be
>Esri Maps for MicroStrategy supported by GIS desktop applications?
>Esri Maps for SAP Business Objects • What can be supported by cost-effective
• Web Application templates web services?
• Operations Dashboard for ArcGIS • What business functions should be
• ArcGIS app for Smartphones and Tablets supported by network services?
• Explorer for ArcGIS • Where will mobile applications improve
• AppStudio for ArcGIS applications business operations?
Training presentation on Boost Productivity Reviewing the available technology
with Explorer for ArcGIS. alternatives and how each performs and
Additional ArcGIS Online SDSwiki content scales within your user environment can
• Software performance provide critical information needed to make
• Information Security the right technology decisions.
System resource selections
Selecting the right technology has a direct
impact on implementation success and
overall return on investment.
GIS data sources
Operations can be supported on local disk or
CD-ROM, shared file servers, geodatabase

servers, imagery, preprocessed map cache, • Provide an efficient way to share data to
or web data sources. support remote client workflows.
• Provide the most efficient way to publish
• Cached map services provide the highest
standard map information products.
performance and scalability.
• Server functionality is available to
• Local data sources support high-
support more advanced user workflows
performance productivity requirements
and services.
with minimum network latency.
• Cost-effective way to leverage GIS
• Remote web services allow connection to
resources to support users throughout
a variety of published data sources, with
the organization and associated user
the drawback of potential bandwidth
communities.
congestion and slow performance.
Intranet applications can access deployed
There are other, more loosely connected,
geoprocessing services managing heavy
architecture solutions that reduce potential
processing tasks in a controlled server
network performance latency and support
environment. Network services can be used
distributed data integration.
to support a variety of web and network
Note: Data selection options will be applications.
addressed in more detail in Lesson 5: GIS
Mobile applications
data administration.
A growing number of GIS operations are
Desktop applications
supported by more loosely connected mobile
The highest level of functionality and GIS solutions.
productivity is provided by local ArcGIS
• ArcGIS technology supports continuous
Desktop applications.
workflow operations that include
• Most professional GIS users and GIS disconnected editing and remote wireless
power users will be more productive with operations.
the ArcGIS Desktop software. • A disconnected architecture solution can
• Desktop applications can be supported significantly reduce infrastructure costs
on the user workstation or through and improve user productivity for some
terminal access to software executed on operational workflows.
central Windows Terminal Server farms.
Best Practice: Leveraging mobile
• Some of the more powerful ArcGIS
services can provide alternative solutions
Desktop software extensions (3D
to support a variety of user workflow
virtualization and heavy Imagery
environments.
processing) perform best on user
workstations with a local data source, Most enterprise GIS solutions include a mix
while most desktop mapping workflows of ArcGIS technology tailored to meet
can be supported more efficiently on a specific business needs. ArcGIS technology
terminal server farm. works together as an integrated system
environment, with each component
Best Practice: Selecting the appropriate
optimized for optimum user productivity.
application deployment strategy can have
a significant impact on user performance,
administrative support, and infrastructure
deployment savings.
Web and network services
ArcGIS Server technologies provide efficient
support for a wide variety of more focused
GIS user workflows.

ArcGIS Enterprise deployment Software technology cycle
Figure 2.35 ArcGIS Enterprise on-premise deployment.
Figure 2.35 shows the ArcGIS Enterprise Figure 2.36 Software technology cycle show the
evolution of a specific software product contributes to
platform architecture. The ArcGIS platform effective GIS operations.
provides a broad range of capabilities to
support your business requirements. Figure 2.36 shows a typical software product
life cycle. Selecting the right technology at
The ArcGIS Enterprise platform includes a the right time is one of the bigger challenges
variety of components that must be for building and maintaining effective
configured and deployed properly to support enterprise GIS operations.
your environment.
Warning: Making the right technology
Selecting the right platform architecture for choice can make or break your success.
meeting your business needs establishes a
framework for successful operations. Technology is changing fast, and new
innovations bring a great deal of promise.
The Capacity Planning Calculator provides a
framework for modeling software Warning: The temptation to select
performance and scalability on available technology based on promise can
hardware and network communication contribute to a painful implementation
technology. The Calculator models what we and failed expectations.
understand about GIS technology patterns Best Practice: Selecting the right
and key workflow performance parameters, technology at the right time can lead to
relationships we can validate with well- optimum success.
defined test benchmarks and operational
experience. For single workflows, the Warning: Replacing aging technology is
Calculator provides a complete software and also important, missing the performance
hardware solution. For Enterprise Design and productivity gains delivered with new
solutions, the Calculator provides workflow technology innovation.
performance targets that can be used in the Software and hardware product life cycles
Enterprise Design. are getting shorter as technology improves
more rapidly each year.
Do your homework.
• Understand your business needs.
• Review available technology
opportunities.
• Complete a design analysis before your
buy.

The Capacity Planning Calculator provides a significant impact on user workflow
framework for modeling software performance, system administration, user
performance and scalability on available support, and infrastructure requirements.
hardware and network communication The following Chapter 2: Software
technology. The Calculator models what we technology selection video provides an
understand about GIS technology patterns overview of GIS Software technology
and key workflow performance parameters, patterns available for use in your Capacity
relationships we can validate with well- Planning Tool analysis.
defined test benchmarks and operational
experience. For single workflows, the CPT Capacity Planning videos
Calculator provides a complete software and The next chapter will discuss Software
hardware solution. For Enterprise Design Performance, providing a much closer look
solutions, the Calculator provides workflow at the software performance parameters and
performance targets that can be used in the baseline performance models in the Capacity
Enterprise Design. Planning Calculator.
Selecting the proper software and
architecture deployment strategy can have a

Software Performance
Fall 2017 GIS Software Performance 41st Edition
This section shares lessons learned about expensive, both in terms of lost services
selecting and building effective GIS design and user productivity. Building a system
solutions that satisfy operational design that addresses capacity planning
performance and scalability needs. needs can improve user productivity and
Software technology allows us to model our reduced implementation risk.
work processes, and provide these models
to computers to optimize user workflow
performance. The complexity of these
Workflow baselines
models, the functions selected to generate
our display, and how application functions
are orchestrated to analyze and present
information processing needs have a
significant impact on computer system
workload and subsequent performance and
scalability.
Figure 3.1 The system architecture design baseline
For many years we focused our system workflow represents a medium load profile distributed
architecture design consulting efforts toward across the baseline software technology components.
identifying and establishing a hardware
infrastructure that would support a standard Workflow baselines provide a foundation for
implementation of Esri software technology. capacity planning. We discussed the
We developed platform sizing models based various GIS software deployment patterns
on consulting experience and customer in Chapter 2 Software Technology. Each
implementation success. We updated our software deployment pattern generates a
sizing models based on relative unique combination of hardware and
performance benchmark testing which network processing loads deployed within a
focused on quantifying changes in critical component architecture that supports the
processing loads introduced with each new system computing environment. Figure 3.1
software release. Today we have a provides an overview of the most important
Capacity Planning Tool that automates our software and network components that
system architecture design analysis impact the system architecture design. We
enabling more refined and accurate will be discussing these components and
performance management. their system configuration strategies
throughout the System Design Strategies
There are examples of GIS deployments wiki. These are the primary components
that do not take advantage of system that work together to support system
architecture design best practices. Systems performance.
are deployed with unresolved performance
issues, and scalability is not well Software technology selection determines
understood. In some cases, performance the software components that will
issues are not identified before the participate in the selected workflow. Each
production system is under critical peak software deployment pattern includes
loads, and often the platform solution or components that are installed on the
network infrastructure fails to meet mission computing system.
performance needs. Resolving performance
issues while in production can be
Software Performance 3.1

• ArcGIS Desktop workstation Custom workflows
deployments. Software includes the
Client and DBMS components. Client
communicates over the network to the
DBMS server.
• ArcGIS Desktop Citrix deployments.
Software includes the terminal Client,
WTS, and the DBMS components.
Terminal client communicates over the
network to the WTS server.
• ArcGIS Server deployments. Software Figure 3.2 Baseline workflow loads are translated to
includes the Client, Web server (Web custom workflow loads by adjusting component
services times by appropriate load factors.
Adaptor), SOC (GIS Server), and DBMS
components. Web client communicates Figure 3.2 shows how we can generate a
over the network to the Web server. custom workflow from a workflow baseline.
The selected software technology baseline
The workflow baseline identifies the medium workflow loads are adjusted by some key
system processing loads for a specific performance parameters. Experience in
software technology pattern. These design, testing, and tuning systems is used
processing loads are expressed as display to establish relative performance
service times and network traffic per display adjustments that can be applied to the
on the various system components. The baseline to generate the custom workflow
load profile (how the load is distributed loads.
across the software components) is fairly
consistent for each software technology Each of the boxes in Figure 3.2 represents
selection. The processing and traffic load key performance parameters that will adjust
intensity will vary within each technology the workflow loads on the system. These
deployment pattern based on display key performance parameters are selected
complexity and some additional key based on decisions you will make in
software configuration parameters we will publishing the map service. Performance
discuss in this chapter. baselines are adjusted based on the
complexity of your data structure or
A workflow baseline is established for each application functions, and display
software technology pattern, identifying a performance is determined by the final
medium complexity load for each of the system architecture design selection. As
system components for that workflow, for you choose your display configuration the
capacity planning purposes. The software workflow processing loads (service times)
baseline loads are established from analysis will be adjusted accordingly.
of benchmark test results and software
deployment experience. Software workflow recipe
The software workflow recipe identifies the
assumptions used to generate the workflow
processing loads. The CPT Calculator is a
tool developed for generating custom
workflow service times (estimated system
loads) translated from a defined software
workflow baseline. The performance factors
in the workflow recipe track the most critical

parameters evaluated during our ArcGIS Server workflow recipe
performance testing.
This chapter will describe each of the critical

recipe components used to establish
software component workflow service times,
and show how to use the CPT Calculator to Figure 3.4 CPT Calculator workflow recipe for an
select appropriate project workflow ArcGIS 10.3 for Server REST service with 2D
performance targets. graphics, Vector only level density (V), medium
display complexity, 100% dynamic (no cached layers
in medium complexity display), 1366x768 map size
There are three workflow recipe formats, (13x7 resolution), PNG24 output format, plus mashup
one for ArcGIS Desktop workflows, one for with a cached basemap (+$$).
ArcGIS Server workflows, and another for
Figure 3.4 shows the CPT cell headings and
Imagery workflows.
selections for an ArcGIS Server workflow.
The ArcGIS Server workflow recipe is
ArcGIS Desktop workflow recipe generated similarly to the GIS workflow.
• The software selection identifies the

appropriate workflow performance
baseline.
• The remaining recipe selections modify
Figure 3.3 CPT Calculator workflow recipe for an
ArcGIS 10.3 for Desktop ArcMap application with 2D the performance baseline to represent
graphics, Vector only (V) display, medium complexity, the selected workflow.
100% dynamic, 1920x1080 map size (19x10
resolution), feature output, with a cached basemap Best Practice: Performance adjustments
(+$$). are selected from look-up tables
established from analysis of benchmark
Figure 3.3 shows the CPT cell headings and
test results.
selections for an ArcGIS Desktop workflow.
The workflow recipe provides metadata
identifying software application design Image service workflow recipe
specifications selected in the CPT
Calculator Software Technology
Performance Factors module.
Figure 3.5 CPT Calculator workflow recipe for an
• The software selection identifies the ArcGIS 10.3 for Server image service using the
mosaic dataset (MosaicDS) for Imagery management,
appropriate workflow performance 2D Graphics display, raster imagery level density (R),
baseline. medium display complexity (Med), 100 percent
• The remaining recipe selections modify dynamic (no cached layers), 1366x768 map size
the performance baseline to represent (13x7 resolution), and JPEG output format.
the selected workflow. Figure 3.5 shows the CPT cell headings and
Best Practice: Performance adjustments selections for an Imagery service workflow.
are selected from look-up tables The imagery workflow recipe is generated
established from analysis of benchmark similarly to the GIS workflow. For an
test results. Imagery workflow, the second word in the
recipe is “Imagery”.
Best Practice: Performance adjustments
are selected from look-up tables
established from analysis of benchmark
test results.

Software technology ArcGIS Pro is a new Desktop application
included with every ArcGIS Desktop license.
selection Professional GIS clients can use ArcGIS
Desktop ArcMap or ArcGIS Pro depending
on which application works best in
supporting their business workflows.
ArcGIS Pro is a popular new ArcGIS

Desktop application with the following
attractive features.
• Provides a modern ribbon menu

interface
• Supports multiple map layouts
Figure 3.6 Software technology selection establishes
• Designed with a Web GIS focus
baseline performance loads for generating custom
workflow performance targets. • Runs side by side with ArcMap
• Provides multi-threaded display
The first step in defining a CPT workflow is performance
to select the software deployment pattern. • Built as a 64-bit Windows application
System performance baselines are CPT ArcGIS Desktop (Desktop) selection
established within the Capacity Planning
Tool for a broad variety of ArcGIS The selection options are ArcMap or Pro.
deployment options. The selection choice is included in the
ArcGIS Desktop Workflow recipe.
The GIS software technology patterns were
introduced in Chapter 2. The CPT
Calculator can generate workflow Graphics 2D/3D performance
processing loads from a variety of software
technology patterns. Your CPT Software
technology selection identifies the
performance baseline used for generating
the selected workflow processing loads.
CPT Software technology pattern
ArcGIS Desktop (ArcMap, Pro)
Figure 3.8 ArcGIS 3D display environments and

ArcGIS Pro are becoming more common for many
business workflows.
Figure 3.8 shows the new display graphics

options (2D, 3D) available in the CPT
Calculator workflow configurations. Dynamic
Figure 3.7 ArcGIS Pro is a new application that can 3D editing and dynamic 3D display
run side by side with ArcMap. CPT Calculator ArcGIS
Desktop workflows now include a Desktop selection
operations are incorporated in the ArcGIS
for ArcMap or Pro. Pro application. Additional 3D display
capabilities are being integrated into
Figure 3.7 CPT Calculator ArcGIS Desktop standard Web client applications. These 3D
includes options for ArcMap or ArcGIS Pro operations place increased graphics
workflows.

processing loads and heavier display traffic • Results above are based on REST 2D
demands on GIS user workflow operations. service.
• Both images are 1366x768 resolutions.
ArcGIS Pro applications take full advantage • JPEG compression works best with
of new graphic processing unit (GPU) raster data (less traffic and faster
capabilities (selected video card becomes display response time).
much more important for display • PNG compression works best with
performance). When deploying ArcGIS Pro vector data (less traffic and faster
in centralized virtual application/desktop display response time).
server environments the server video card
becomes the primary capacity limitation PNG24 is CPT Calculator default for vector
when configuring your server environment. density (vector only and VPortal).
Dynamic 3D environments can also be
much more compute intensive than • Supports transparent overlay.
traditional 2D environments, both in terms of • Compresses common pixel values.
the complexity of the data model (number of
features per layer) and the application 3D Best Practice: Use PNG8 for vector
dynamic processing requirements. Dynamic business layers when higher color depth
3D display is where technology is moving – is not required.
although initial operations may challenge
existing server processing and network JPEG is default for raster imagery density
traffic capacity limitations. (Raster image or RPortal).
CPT Display Graphics selection • Common compression across different

pixel values.
Data density makes a difference
Best Practice: Use JPEG output when
imagery layers are included in the map
display.
CPT Calculator Density/Portal selections
Figure 3.10 shows a web service shared in

a Portal workflow. Web services are
registered with Portal when services are
included in a Web Map shared within the
Figure 3.9 Data density makes a difference. Portal organization or added as a registered
service. Portal registered workflows incur
Map density has a direct impact on display additional processing loads on the Portal
compression. When using a PNG output Web server.
format, solid color areas compress better
than areas with rapid color change. Less
display compression means more display
traffic, which requires more time to deliver to
the client.
Figure 3.9 shows a vector and raster output

of the same map along with the different Figure 3.10 Registered Portal Workflow. ArcGIS
traffic and travel time for each output format. Server workflows can be registered with Portal for
Density of the data will make a difference in content management and collaboration.
processing and output image compression. CPT ArcGIS density/portal selection

The density selection modifies traffic and Tradeoff between map display quality
rendering time processing loads. Density and user performance
selection applies to all Server workflows,
and includes a heavier Web server load
when services are registered with a Portal
for ArcGIS service.
CPT Calculator density selections include

the following:
• Vector only
• Raster image
• VPortal
• RPortal
Figure 3.11 Classic dynamic mapping trade-off.
Vector only (V) or VPortal are used when Figure 3.11 shares the tradeoff between
display is limited to vector feature types. quality and speed when publishing a
Raster image (R) or RPortal are used when dynamic map service. The functions and
the display includes an imagery layer. Portal analysis included in the map will impact
should be included in the workflow density display performance. Higher quality maps
selection when the Web service is require more dynamic processing, while
registered with a Portal for ArcGIS Web simple maps can be rendered much faster.
server.
Both high-quality and simple maps can
Display complexity provide very similar information, but may
show very different performance.
Display complexity is an estimate of how
much processing a computer system must • High-quality maps often include heavy
do to complete a unit of work. Workflow functions such as shaded relief,
display complexity includes a broad range transparent layers, and dynamic Maplex
of software and data design factors. The labeling which results in slow
complexity determination used for initial performance.
capacity planning is often a rough estimate • Simple maps do not require heavy
related to a standard (medium complexity) functions, and the example shows low-
baseline workflow processing load profile. resolution relief, solid colors, and simple
There will be opportunities to measure the annotation which enables fast display
display complexity during initial publishing performance.
and deployment of the map service. Best Practice: High-quality map features
served as a cached basemap perform
very well.

GIS user performance expectations are introduced, it will be increasingly
important to plan, set performance
milestones, and manage compliance during
system deployment.
At the same time, there are more

opportunities than ever before to reduce the
risk of deploying systems that do not meet
your performance needs.
Map display complexity
GIS provides users with a geographic view

of their business environment, and for many
Figure 3.12 User performance expectations have GIS workflows, the map display is used as a
changed over the last 10 years primarily due to faster
hardware processing technology. Heavier processing primary spatial information resource.
loads can now be supported without impacting user Warning: Not all map displays are
productivity.
created equal.
Figure 3.12 shows how user performance
CPT Display complexity selection
expectations have changed along with the
technology. GIS project efforts can be
completed in less than half the time it took Display complexity is used to establish
just 10 years ago. GIS professionals used to workflow performance targets; selection
wait on computers to do their work – today identifies the display processing load
computers are waiting for users to review relative to a medium complexity baseline
and provide feedback. workflow.
User display performance expectations in High performance GIS mapping services

2005 were around three seconds—a
challenge for medium map displays viewed
in the computer room. The same map
service today can be rendered in less than
0.3 seconds. These performance
improvements open opportunities for:
• More complex dynamic map services.

• Deployment of ArcGIS Server on easier-
to-manage virtual server environments.
• Deployment of web services on a
hosted cloud computing environment.
• The possibility of much richer dynamic
services that employ more sophisticated
statistical analysis or network routing
algorithms (two to three times the
complexity of current GIS workflow
baselines).
Figure 3.13 GIS (vector) workflow display complexity.
These opportunities will introduce new Intensity of the workflow processing loads will depend
on the display complexity.
challenges. As heavier processing options

Figure 3.13 shows the display performance loads than a medium workflow. Heavy
of twelve levels of complexity for a typical workflows represent richer display
GIS workflow. Software technology services used more often with
selection establishes how the workflow introduction of faster hardware
processing loads are distributed across the processing capabilities.
software components. The complexity • 2xMedium: Twice the loads of a medium
selection provides your estimate of the workflow.
intensity of the processing loads. This • 3xMedium: Three times the loads of a
complexity estimate considers hundreds of medium workflow.
factors that are often observable in the map • 4xMedium: Four times the loads of a
display. medium workflow.
• 5xMedium: Five times the loads of a
The first step in publishing a map service is medium workflow.
to create a map document representing the • 6xMedium: Six times the loads of a
geographic information product you wish to medium workflow.
publish. The number of layers in the display, • 7xMedium: Seven times the loads of a
the number of features in each layer, the medium workflow.
types of processing functions required to • 8xMedium: Eight times the loads of a
render the map display, and the complexity medium workflow.
of the data model will contribute to display • 9xMedium: Nine times the loads of a
complexity. medium workflow.
• 10xMedium: Ten times the loads of a
• The software functions used to generate medium workflow.
map documents can represent the
heaviest processing loads within the The complexity of the authored map
user workflow. Other functions used in document will be a primary factor in
the analysis besides what is required to determining map service performance and
produce the display should also be scalability. Light maps are rendered three
considered. times faster than heavy maps, and six times
• The types of functions, data source faster than the much heavier 3x Medium
format, and design of the user display complexity maps.
can make a big difference on the level of
processing and network loads required The following are some best practices for
to support a GIS user workflow. authoring high-performance web maps. Use
these recommendations as a guide to build
The CPT Calculator provides twelve choices a map that will perform within your
for complexity. performance budgets.
• Light: Fifty percent of the loads of a Only show relevant data.

medium workflow. Use for very simple • Start simple.
map displays with a minimum number of • Use field visibility.
focused layers and no heavy functions. Use scale dependencies.
• Medium: Standard workflow baseline, • Display appropriate data for the given
suitable for most dynamic mapping scale.
workflows that follow standard best • Display the same number of features at
practices for high-performance mapping. all scales.
• Heavy: Heavy functions, larger number Select the appropriate point representation.
of display layers, or more features per • Use single-layer simple or character
layer require heavier processing loads markers.
for this workflow. Fifty percent more • Use EMF instead of bitmaps.

• Use integer fields for symbol values. • Limit the maximum number of records
• Avoid halos, complex shapes, and returned per request (mosaic dataset
masking. only).
Select the appropriate lines and polygons. • Select the appropriate metadata level.
• Use the ESRI Optimized style. o Basic, full, or none
• Avoid cartographic lines and polygon • Set the appropriate allowed fields
outlines. (mosaic dataset only).
Use appropriate text and labeling.
Download parameters:
• Use annotation instead of labels.
• Use indexed fields. • Select appropriate output and virtual
• Use label and feature conflict weights directories.
sparingly. • Identify supported image return type.
• Avoid special effects (fill patterns, halos,
Best Practice: Use the map publishing
callouts, backgrounds). analysis tool to identify potential
• Avoid very large text size (60+ pts).
performance tuning opportunities.
• Avoid Maplex for dynamic labeling
Warnings identify opportunities for
(avoid over-use). performance tuning and show help
documentation for adjusting complexity
of your map service.
High performance Imagery services
Use MXDPerfStat to measure map
Similar display complexity performance display complexity
variations apply to imagery workflows. The
following blog posted by David Johnson
from the Esri Prototype Lab Sales Support
team shares recommendations for
Designing and Optimizing Image Services
for High-Performance Analysis. Following
are some best practices for authoring high-
performance imagery products, in order to
keep the image service time within
performance budgets.
Image parameters:
• Limit the maximum image size per
request. Figure 3.14 MXDPerfstat is an ArcScript that
• Limit the maximum number of rasters measures render performance for a defined map
per mosaic. extent. Results show render time for map layers at a
range of scales at a select geographic location.
• Select the optimum resampling method.
o Discrete data (Nearest Neighbor The [MXDperfstat tool] identifies display
or Majority) refresh times at multiple scales, shows layer
o Continuous data (Bilinear refresh times for each map scale, identifies
Interpolation or Cubic data sources for individual layers, and
Convolution) provides layer performance statistics [such
• Use the optimum compression method. as number of features, vectors, labeling,
• Set the optimum compression quality. and breaks out display time for several key
• Set the optimum mosaic method. rendering phases (geography, graphics,
cursor, and database)].
Catalog parameters:

Figure 3.14 shows some sample required to produce an average display. For
MXDPerfstat results run on a file GIS mapping workflows, most of the display
geodatabase dataset. Results are rendering time is consumed in creating the
summarized for display purposes. dynamic map product. Figure 3.15 provides
an engineering chart that shows typical
Best Practice: Caching the basemap
MXDPerfStat rendering times for light,
would reduce display complexity by 50
medium, and heavy map displays as a
percent.
function of the host platform.
MXDperfstat is an excellent tool for
• Medium map display rendering time with
measuring map document display
a high-performance 2017 platform
performance, since it lists layer statistics
(Xeon E5-1270v5 3600 MHz
(render time, features, edges, projection,
processors) would be about 0.334 sec.
etc.) for each layer included within a
• The same map display rendering time
complete series of map scales. The
measured on a 2009 Intel Core i7 2933
measured results can be used for
workstation would be over 0.79 sec.
evaluating map display complexity and
tuning your map document for optimum
display performance. Rendering time is a function of display
complexity and platform performance.
Best Practice: Measure display render
performance during initial map authoring • Platform performance is represented by
to validate compliance with design vendor-published platform SPEC
performance targets. benchmark results.
• Faster computer platforms will render
Once you see the layer processing time and the map in much less time than slower
the performance metrics, the layers with platforms.
display problems are usually easy to spot.
Best Practice: Measure display
performance during map publishing to
verify within design performance targets.
Defining display complexity
Dynamic map display results shown in
the chart were rendered from a local file
geodatabase data source using the
ArcGIS Server Service editor preview
tool.
Warning: Display complexity is a relative
term representing a variety of system
performance variables. Workflow display
complexity is important for capacity
planning since it directly contributes to
system performance and scalability.
ArcGIS Desktop provides a map publishing

Service Editor with specific performance
Figure 3.15 Display complexity is defined by the
processing time required to render a map display from tuning tools. Since ArcGIS 10.1, the
a local file geodatabase. Measured time can be Analyze and Preview tools appear at the top
normalized by adjusting for relative processor of the Service Editor during the map
performance. publishing process.
Display complexity is an estimate of the
average amount of computer processing

[Map service publishing tools] are described • Information messages identify additional
in the ArcGIS online documentation considerations for performance tuning.
available on the ArcGIS Resource Center.
ArcGIS help Warning Messages section
includes a long list of potential performance
Use map publishing tools to measure problems. This list of error, warning, and
display complexity information messages can be used for map
document tuning.
ArcGIS Desktop provides a map publishing
Service Editor with two very important 3. Preview will let you zoom and pan around
performance tuning tools. Starting with your map to check performance.
ArcGIS 10.1, the Analyze and Preview tools
appear at the top of the Service Editor The Preview tool can be used to measure
during the map publishing process. The display complexity. The tool identifies the
following demonstration shows how to use map rendering time output for the current
the ArcGIS Desktop publishing analysis and map display. The map render time is
preview tools for managing map display provided by the [ArcGIS Desktop map
complexity. publishing preview tool].
Map service publishing tools service editor • Rendering performance is provided at

demonstration the top of the display.
• Set the planning image output format for
The warnings identified by the map your service.
publishing analyze function provide • Adjust the preview display image size to
opportunities for performance tuning. Steps match the average client display
to create a high-performance map. resolution.
• Move around the more dense areas to
1. Design and create a map document in measure display complexity.
ArcMap. • Review display performance at each
scale to provide a consistent user
2. Analyze your ArcMap document by experience.
clicking Analyze on the map publishing • If performance is not acceptable, review
Service Editor. the Analyze function warnings for
additional opportunities to improve
The Analyze function will identify error performance.
messages, warning messages, and Best Practice: Use the map publishing
information messages. tools during the authoring phase to
evaluate map display complexity. Render
• Error messages identify issues you must time targets can be established to author
fix before publishing the map. a published service within your
• Warning messages identify issues that performance budget. Early performance
may affect drawing performance and validation can reduce implementation
appearance. risk.

Take advantage of caching
Map cache can reduce the number of (%DataCache)
dynamic display layers
Figure 3.16 Map displays are created one layer at a Figure 3.17 There are clear performance advantages
time. Parallel service requests can improve render when using a pre-processed tiled map cache for your
performance, but will increase network traffic and may basemap layers.
not improve user productivity.
Figure 1-17 shows the advantage of map
Figure 3.16 shows the layered rendering caching. Map tiles are processed once to
process for a dynamic map service. GIS create the cache (cachingtools provide
displays are created one layer at a time, background geoprocessing service
similar to the procedure geographers would instances for creating the map cache).
follow to lay out a map display on a table Client display request for cached map tiles
using mylar sheets. is routed to get the requested tiles, which
are then downloaded to the client for display
• The technology has changed, but the (negligible server resources required for
procedure for building a map is much fetching the tiles). Once downloaded to the
the same. map display, the tiles are cached in the
• Maps with a few layers require less client browser. The second request for the
processing than maps with many layers. tile comes from browser cache (not from the
• The layer complexity (number of server). Tiles can also be cached at
features, edges, symbology, tasks, etc.) locations between the server and browser
impacts rendering time for each layer. cache, and a copy of the closest available
• Building a map display renders one tile is delivered to the client. ArcGIS Online
layer at a time, joining the features uses a third-party caching solution to
(points, polygons, lines) in each layer distribute cached basemap tiles at global
sequentially, one on top of the other, locations throughout the internet, providing
until the final display is complete. a very scalable tile delivery configuration
where a majority of tile requests are
Parallel query displays can be published serviced before reaching the GIS Server
with ArcGIS Server technology—but is the location. The location of the requested map
performance gain worth the use of extra cache tiles is the primary factor contributing
shared infrastructure resources? to cached map display response time.
ArcGIS Server provides a variety of ways to

maintain and support pre-processed maps,
and to organize the map files in a map
cache structure optimized for map
publishing.

A map cache reduces the number of services presented as a transparent
dynamic display layers (less processing image [mash-up (business layers)] over
load). a map cache base layer.
• Dynamic map services are important
Preprocessed basemap images are for geographic analysis, editing, and
available in optimized map cache format. geoprocessing functions, which
require access to point, polygon, and
• Maps can include dynamic and cached line features rendered in a dynamic
layers. map display.
o Operational layers generated • Map cache tiles provide an optimum
from a dynamic data source. basemap foundation layer,
o Basemaps delivered from a combining high-quality map
preprocessed map cache (static visualization with high display
layers). performance.
• Client display combines operational • A mash-up of dynamic operational
layer rendered over cached basemap layers over high-quality base map
(mash-up/overlay). reference layer delivers the optimum
• Cached basemaps are available from combination of quality and
ArcGIS Online. performance.
Data classification is application-specific. ArcGIS Online Basemaps
• Dynamic layers provide the best choice ArcGIS Online shares basemaps and
for most rapidly changing data. reference layers freely available to anyone
Examples include roads showing snow with Internet access. Published basemap
depth or electrical network showing services include World Imagery, World
latest posted work orders. Street Map, World Topographic, Ocean
• Static layers provide the best choice for Basemap, and more. Data appliance for
slowly changing data. Examples include ArcGIS is an Esri product designed and
land use/land cover, road networks, and optimized for use with ArcGIS Server to
basemap data. deliver secure mapping applications behind
an organization's firewall.
The quality of the fully cached map can be
much higher than the medium dynamic CPT ArcGIS percent data cache
display (and map publishing performance (%DataCache) selection
still the same), the difference is that the fully
cached map processing was completed
The map cache setting identifies the
before posting on the website, and the final
percentage of display layers that will be pre-
processing time for the cached map tiles is
processed into a tiled map cache. The
minimal.
percent dynamic is calculated (1-
Best Practice: The optimum web %DataCache) and the %Dyn percentage is
mapping display combines dynamic map included in the Workflow recipe.

Warning: The required amount of traffic
Selecting the best map resolution per display can have a significant impact
on user performance over lower
bandwidth.
Note: Server processing loads may also
increase with higher resolution displays
(higher resolution can result in larger map
extent, increasing the number of features
rendered for each client display).
CPT Resolution selection
Calculator resolution selection identifies the

map output display size. More traffic is
required when publishing larger map
Figure 3.18 Selecting the right image resolution. resolution.
The resolution of the map display output
can be an important consideration for GIS
user workflows. A high-resolution display Selecting the best output format
with full-screen map size is important for
users who work with map displays over long
periods throughout the day. A display
resolution of 1920x1080 or higher may be
common for GIS power users or data
maintenance (GIS editor) workflows. Rich
Internet applications (Flex and Silverlight)
may also present higher resolution map
displays.
Figure 3.18 shows the variation in traffic and

response time for a 600x400 and 1366x768
pixel display resolution of the same map. Figure 3.19 Selecting the right output format.
Remote client performance with a 1366x768 Web mapping services produce map
display on a dedicated T-1 connection images or features that are sent to the client
would be over 3 times slower than a local browser for display. Figure 3.19 shows the
user display. A display resolution of variation in traffic and display processing
600x400 is a common size for imbedded time adjustments for the different output
web browser map displays. Remote display formats.
performance over a 56Kbps line with a
small map is less than 10 seconds.
• With image services, each user request
will generate a new map image that
Web mapping services produce map must be delivered to the client browser.
images that are sent to the client browser • The selected image type can have a
for display. Each user request will generate direct impact on the volume of network
a new map image that must be delivered to traffic. Lighter images require less
the client browser. The size of the output display traffic and heavier images
image varies directly with the number of require more display traffic.
pixels, so higher resolution images generate • The required amount of traffic per
much higher client traffic loads. display can have a significant impact on
user performance over lower bandwidth.

Selecting the best vector data source
Image output selection observations: format
• Vector-only images compress better

than images that include a Digital Ortho
raster layer.
• JPEG image types provide the most
consistent compression, with minimum
variation between raster and vector
images.
• PNG images compress much better with
vector data than with raster—PNG
supports transparencies and is the
default ArcGIS Server output format.
• PDF is a heavier output format used for
high-quality map plotting. Figure 3.20 Several output service formats are
available for map publishing. Select the format that
Best Practice: Select the data format that applies to your published service. Calculator will use
provides the best display performance the output look-up table to adjust traffic and loads
while meeting your business based on your selection.
requirements. Figure 3.20 shows the application service
CPT Output selection time and traffic adjustments used by the
CPT to adjust for vector data source
Several output service formats are available selection. How you store and manage your
for map publishing. Select the format that data makes a difference in performance.
applies to your published service. Calculator
will use the output look-up table to adjust • Maintenance (production) Enterprise
traffic and loads based on your selection. Geodatabase (SDE_DBMS)is needed
for data integrity (some performance
overhead).
Data source selection • Simple feature Enterprise Geodatabase
(publishing database) provides
Data source is selected for each user improved performance over production
workflow on both the CPT Calculator and database.
the CPT Design tabs. Data source selection • A file geodatabase performs well in a
is configured separate from the workflow static, read-only environment.
service times, so a defined project or • Larger shapefile data formats incur
standard workflow can be used to create increased traffic and performance
multiple use cases on the CPT Design tab overhead.
each with a different data source. On the
Note: Data source management and
CPT Design tab, each workflow selection
deployment options will be discussed in
(row) in the requirements module will have a
more detail in Lesson 5: GIS data
defined data source.
administration.
CPT Vector storage format selection
Several vector data source formats are

available for your selection. Calculator will
use a look-up table to adjust output traffic
and loads based on your data format
selection.

Selecting the best imagery storage • JPEG 2000 provides good storage
format traffic compression but incurs high
performance overhead.
• LizardTech MrSID compression reduces
storage traffic with moderate
performance overhead.
• Intergraph ECW format reduces storage
traffic with moderate performance
overhead.
• ERDAS IMG format requires moderate
performance overhead and high storage
traffic.
• Storing imagery in an Enterprise
geodatabase reduces storage traffic and
provides good performance. This format
Figure 3.21 CPT Imagery storage formats.
is read-only, and is no longer a
preferred Imagery storage option.
Figure 3.21 shows the application service
time and traffic adjustments used by the
There are many factors that drive how you
CPT to adjust for imagery data source
may store your imagery. The CPT provides
format selections. How you store your
a limited sample of storage options, and
imagery makes a difference in performance.
performance metrics can change with each
software release. Performance data
• TIFF uncompressed provides the best
provided is based on preliminary test
performance, but requires high storage
results. Much better information on
traffic.
performance of ArcGIS Server Image
• TIFF LZW compression results in
Services should be available as the ArcGIS
reasonable performance overhead with
technology matures.
high storage traffic.
- LZW provides over 60 percent
CPT Imagery storage format selection
compression with 8-bit imagery (smaller
storage).
- LZW provides over 20 percent larger Several imagery data source formats are
files with 16-bit imagery (more storage). available for your selection. Calculator will
• TIFF JPEG compression results in good use a look-up table to adjust output traffic
storage traffic reduction with reasonable and loads based on your data format
performance overhead. selection.

Custom workflow Standard workflows are included on the
CPT Workflow tab. You can select Standard
processing loads Workflows on the Calculator tab for single
workflow sizing.
Custom workflow performance loads are
generated from baseline workflows.
Selected software performance factors are Software performance
applied to generate custom baseline service
times. CPT Calculator workflows are
summary
generated from performance benchmark
baselines. Experience suggests we can do a better job
selecting and building better software
Custom workflow service times created on solutions. Understanding software
the Calculator tab are copied to CPT performance can reduce implementation
Workflow tab. Workflow service times can risk and save customer time and money.
then be copied and included in Project Projects can be delivered within project
Workflows. Calculator workflows provide a cost, time, and performance budgets.
source for both standard and custom
workflow performance targets. The next section will take a closer look at
ArcGIS Server software performance.
Best Practice: Software technology
baseline service time, traffic loads, and
relative performance adjustments are CPT Capacity Planning videos
derived from test benchmarks.
CPT Standard workflow selections

Server Software Performance
Fall 2017 GIS Server Software Performance 41st Edition
ArcGIS 10.5 software release includes After this chapter, you will be ready to
ArcGIS Server as a component of the explore how network communications,
ArcGIS Enterprise software license. ArcGIS platform architecture, and server platform
Enterprise provides a variety of ArcGIS selection complete the system design
Server capabilities, supported by a fully components required for implementation
integrated services architecture including success.
portal content management, ArcGIS Server
processing capabilities, and a variety of
optimized data source configurations. This
ArcGIS Enterprise software
chapter will focus on the ArcGIS Server architecture
component configuration strategies for
optimum system performance and
throughput capacity.
ArcGIS Server performance and capacity

depends on a proper service configuration.
Once you have a basic understanding of
software performance, you are ready to
better understand how to properly configure
services for optimum server utilization.
The previous chapter on software Figure 4.1 ArcGIS Enterprise software components.
performance introduced the concepts
required to recognize the primary software Figure 4.1 shows the ArcGIS Enterprise
components and data structure that architecture components.
contributes to system performance. This
chapter looks more closely at the specific • ArcGIS Server provides the layers,
decisions map service authors and system services, and horsepower required to
administrations must make in properly power your Web GIS. ArcGIS Server
configuring services for optimum capacity includes capabilities for traditional
and throughput. Standard and Advanced capabilities
with additional licensing for a variety of
ArcGIS Enterprise platform selection and integrated ArcGIS Server roles.
licensing assumes you understand how to • Portal provides the web interface and
select the proper service configurations for infrastructure content that supports a
optimum throughput and performance. user’s interaction and overall experience
Selecting the proper hardware and licensing with your Web GIS.
establishes the potential for optimum • ArcGIS Data Store provides the Esri
system performance; proper service managed data repository that stores the
configuration is essential to realize this spatial content that has been shared to
potential. This chapter provides you with the Portal. Configurations include the
guidelines and best practices you need to Relational, Tile Cache, and
achieve optimum performance from your Spatiotemporal Data Stores. ArcGIS
ArcGIS Server system deployment. Enterprise also integrates with
traditional data sources (Enterprise
Server Software Performance 4.1

Geodatabase, tabular business ArcGIS Server key component functions.
systems, etc).
• Web Adaptor is an Esri built software • Web gateway: Web server software
load balancer that appropriately directs installed with the web adaptor.
network traffic and serves as a reverse • Web adaptor: Provides site-aware load
proxy for Web GIS access. distribution, reverse proxy, and failover.
• GIS server: Provides ArcGIS software
Note: Further discussion on ArcGIS for site and cluster service instance
Enterprise architecture is provided in management.
Chapter 9: GIS product architecture. • ConfigStore: File share that stores site
service configuration parameters.
• SvrDirectories: File share that stores
site directories.
ArcGIS Server site • Data source: Common data source for
configurations site services.
The standard ArcGIS Server site supports Warning: If you lose the ConfigStore or
multiple machines deployed within a single data source, you lose the site.
site configuration, with each machine ConfigStore backup and recovery was
supporting a common set of published introduced with the 10.2 release.
service configurations. ArcGIS 10.5 retains
a legacy option for multiple clusters within a ArcGIS Server site capabilities can be
single site, with each cluster of machines modified for improved performance and
able to support a different set of published scalability.
service configurations. Multiple-cluster site
configurations include an internal cluster Single-cluster mode (introduced with
aware load balancing service. ArcGIS 10.3.1 release)
ArcGIS Server component functions
Figure 4.2 ArcGIS Server is a component

environment designed for easy development,
deployment, and site- aware scalable architecture.
Figure 4.3 ArcGIS Server site single-cluster mode
Figure 4.2 provides an overview of the key with no internal cluster aware load balancing.
web components used to define the ArcGIS
Server architecture. Figure 4.3 shows a view of the ArcGIS
Server single-cluster architecture. ArcGIS
Server single-cluster site configuration

provides improved throughput and currently using single cluster mode.
scalability. Single cluster sites already using single
cluster mode and sites with multiple
• ArcGIS Server internal site load clusters at earlier releases will have
balancing disabled. their settings preserved when
• Reduces internal site network traffic. upgrading.
• Improves site multiple machine peak • When using single cluster mode, all GIS
throughput (improved scalability). servers in the site must participate in a
single cluster. Multiple clusters cannot
Best Practice: Single-cluster mode is the exist.
default configuration with the ArcGIS
10.4+ releases. Read-only mode (introduced with the
ArcGIS 10.4 release)
Single-cluster mode selection
Figure 4.4 ArcGIS Server Administrator Directory

single-cluster mode selection. Figure 4.5 ArcGIS Server site read-only configuration,
with copy of ConfigStore and SvrDirectories cached
The ArcGIS Server Administrator Directory on each server machine.
is used to set the single cluster mode
deployment configuration. Figure 4.4 shows Figure 4.5 shows a view of the ArcGIS
the single cluster mode selection. Single Server read-only architecture. ArcGIS
cluster mode is active when the single Server read-only mode improves site
cluster mode setting is true. stability by reducing dependency on
connection to the ConfigStore network
• At 10.4 and 10.5, ArcGIS Server share.
defaults to single cluster mode for new
installations. This mode does not • ConfigStore cached on each local GIS
include load balancing between Server machine.
machines in the cluster. This reduces • Service publishing and changes are not
network traffic between machines in the supported.
site, reduces load on your network, and • Allows read operations in degraded
helps improve monitoring of GIS servers mode.
in your site.
• When upgrading from earlier releases, Operates in degraded mode with lost
the single cluster mode will be enabled connection to file share
for sites with a single cluster that are not

• Geoprocessing services will not work. ArcGIS Server Terminology
• Consumption of cached services will not
work unless they are highly available.
• Exporting a map or image service using
a URL is not possible.
Read-only mode selection
Figure 4.7 ArcGIS Server includes instances,

processes, and threads. Understanding this
terminology is important in properly configuring the
server.
Instances, processes, and threads

determine what software will be deployed to
Figure 4.6 ArcGIS Server read-only site mode
selection. leverage the available hardware processing
resources. Figure 4.7 shows the instances,
The ArcGIS Server Manager is used to processes and threads within the ArcGIS
enable the read-only site mode Server configuration.
configuration. Figure 4.6 shows the read-
only mode selection. ArcGIS Server Site ArcGIS Server Site
Administrator can enable the read-only
mode on the ArcGIS Server Manager site
An ArcGIS Server site refers to one or more
settings tab.
GIS Servers with a common configstore.
Beginning at 10.4, ArcGIS Server provides
• ArcGIS Server licensing requirements
a read-only mode that allows you to control
are established by the highest service
changes to the site. Enabling this mode
instance capability deployed within the
prohibits the publishing of new services and
site.
blocks most administrative operations. Your
• Licensing is based on customer number
existing services function as they previously
and computer datacenter location
did. For example, you can still edit data in
• Single ArcGIS Server site can be
feature services when the site is in read-
configured to support multiple machine
only mode.
clusters.
• All machines within a cluster have the
same service configurations.
Server process
A server process refers to a software server

executable.
• ArcGIS GIS Server (SOC.exe) process

• Web server process

• Web Adaptor process Service instance configuration
Process threads Each published service has a defined

service instance configuration which
The term “Process” refers to the package of represents how that service will be deployed
program executables contained within each through the software stack. Service instance
deployed SOC (each deployed SOC configurations are assigned to service
contains one copy of the ArcGIS program clusters, and multiple service clusters can
executables). Some server processes are be assigned within a server site.
multi-threaded (low isolation), and allow
multiple service requests to be processed at • Each GIS Server machine within a
the same time using a single copy of service cluster will deploy the same
process executables. service configurations.
• ArcGIS Server Site is defined by a
• ArcSOC.exe processes can be common (shared) Configuration Store
configured to support multiple service (ConfigStore). The ConfigStore is a file
threads. that contains the cluster assignments
and associated service configurations
The term “thread” is used at the hardware for the ArcGIS Server site.
level to represent an on-chip location to
park a service instance close to the core for
processing (a hyper-threaded core can Batch process service instance
reduce the time required for context configuration (geoprocessing services)
switching when multiple concurrent
execution threads are sharing a single core
processor; objective is to better utilize
available core processing cycles).
Service instance
Service instance is a process thread made

available for servicing a map request. The
service configuration identifies the minimum
and maximum number of service instances Figure 4.9 Batch process load profile
that will be deployed by ArcGIS Server to
satisfy inbound web service requests. The brown line in Figure 4.9 shows the
maximum host platform utilization and
• Minimum number of specified service system throughput in displays per minute
instances for each service will be (DPM) for a series of ArcGIS Server batch
deployed during server startup. process service configuration instance
• Additional service instances will be settings. The bars show host platform
deployed by the service manager based service time (colored tier) and service wait
on the number of concurrent service times (white wait times are due to shared
request arrivals, and can expand the use of the available core processor). The
number of service instances up to the host platform has 4 core; the four core
maximum specified for each service processors are shared resources used to
instance configuration. execute the deployed service instances.
The first bar represents a batch process

configuration with one (1) service instances.

Peak service load is roughly 25 percent host instance settings to optimize site
platform utilization with peak throughput of performance. If you need to work on the
190 DPM, well below the maximum host site cluster during batch geoprocessing,
platform throughput capacity (display you may want to configure less than N+1
service time is normally not important for instances to allow for an administrative
batch process loads – more attention is processing reserve.
given to how long the total batch job will
run). Optimum service instance configuration:
The fifth bar represents a service • Minimum instance = Zero (0) for long
configuration with five (5) service instances. running batch processes, One (1) for
Host platform approaches 100 percent lighter network services.
utilization with minimum increase in batch o ArcSOC process start-up time is
run time. Display response time (including often short compared to overall
total batch service run time) will increase batch load.
linearly once server is operating over 99 o Reserves memory and capacity
percent utilization. A service configuration space for other active map
with ten (10) service instances would take services.
twice as long to complete each batch job. o Service instances can grow up to
Peak throughput is normally reached at N+1 the max instance setting during
service instances (host platform core + 1). peak loads.
Increasing the number of service instances • Maximum instance = Provide one more
will only increase batch processing times – batch instance than available server
it is better to queue up processes and machine cores (N+1 instances where N
complete jobs sequentially than try to run = number of server cores).
them all at the same time. o Provides optimum throughput
during peak service loads.
Batch process minimum and maximum o Minimizes batch runtime
instance settings are different than map performance during peak loads.
services. Batch processes consume a • Limiting maximum number of service
server processor core. instances can be used to limit platform
core processor utilization.
• Process queue provides consistent o Single service instance will use
processing load. less than one core processor
• Minimum number of service instances during maximum service loads.
should be assigned to optimize
processing performance. Many ArcGIS services perform as batch
processes.
Service configuration min/max instance
settings impact server performance. • Most heavy geoprocessing services
• Batch address locator services
• More service instances provide more • Map caching services
throughput when available instances are • Map printing services
less than available processor cores. • Heavy network routing services
• More service instances increase batch • Imagery processing services
runtime when available instances are
more than available processor cores. A batch process can be any service that
runs for a long time without user input.
Best Practice: Set map service Batch processes include geoprocessing,
configuration minimum and maximum map cache generation, system backup, map

production script, replication services, or Web service instance configuration
any other calculation that may take longer (concurrent users)
than 60 seconds to complete without user
input. This type of service should be
initiated from the client application as a work
request, and delivered to a queue for
sequential processing by a batch process
service.
Best Practice: Batch processing can be

the most efficient way to handle heavy
processing loads.
Warning: For handling a heavy batch
service, the maximum instances should Figure 4.10 Web mapping service instance load
be a small number to protect site profile.
resources. A single, heavy batch service
can consume a server core for the length The brown line in Figure 4.10 shows the
of the service process time. Four maximum host platform utilization and
concurrent batch requests on a 4-core system throughput in displays per minute
server can consume all available host (DPM) for a series of ArcGIS Server service
processing resources. configuration instance settings responding
to random Web service requests. The bars
Batch process workflow can be configured show GIS Server machine service time
in the Capacity Planning Tool by setting the (colored tier) and service queue times
workflow minimum think time = 0. Batch (white processing queue times due to
workflows can be used on the CPT Design random arrival of service requests). The
tab to reserve system resources for batch server machine has 4-core; the four core
processing during heavy loads. You will processors are shared resources used to
need to use the RESET ADJUST function to execute the deployed service instances.
calculate batch service productivity for the
maximum server load. The first bar represents a service
configuration with two (2) service instances.
Peak service load is limited to less than 38
CPT ArcGIS Server batch process instance percent host platform utilization with peak
configuration throughput of 287 DPM and display
response time just over 0.40 seconds, well
CPT Calculator and Design tabs used to below the maximum host platform
demonstrate the optimum number of batch throughput capacity. The eighth bar
process service instances. represents a service configuration with
sixteen (16) service instances. Host
platform is over 92 percent utilization with a
display response time over 1.35 seconds.
Average display response time continues to
increase as additional service instances are
deployed with minimum increase in
throughput. Peak throughput is normally
reached at about 3 to 5 service instances
per host platform core (12 to 20 service
instances on a 4-core host platform).
Increasing the number of service instances

beyond this level will only increase the o Service instances can grow up to
average display response time with minimal the maximum instance setting
throughput gain. during peak loads.
• Maximum instance = Three to five
Random arrival distribution reduces peak instances per server machine core.
throughput per service instance. o Provides optimum throughput
during peak service loads.
• Peaks and valleys in the arrival o Optimizes display response time
distribution place varying demands on performance during peak loads.
server processing loads.
• Sufficient number of instances must be CPT Design map service instance
available for service assignment to configuration
achieve peak throughput values.
CPT Design tab can be used to
Service configuration min/max instance demonstrate the optimum number of Map
settings impact server performance. service instances.
• During light server loads, more service Map Service instance

instances can increase server peak
throughput. configuration strategies
• During heavy server loads, more service
instances have minimum impact on Minimum and maximum service instances
peak throughput. are identified when publishing a map
• More service instances increase display service. It is important to identify the proper
response time. instance configuration for each map service
deployment. Proper service instance
For a popular service, the maximum service configurations depend on the expected
instances should be large enough to take peak service demands and the server
advantage of the full site capacity. Full site machine processor core configuration.
capacity may require 3-5 service instances
per core. A site with 4-core GIS Server
machines would require a maximum service Pooled service model (service
instance setting of 16—the recommended instance settings)
instance setting for a popular map service.
Best Practice: Set map service

configuration minimum and maximum
instance settings to optimize site
performance.
Optimum Web mapping service instance

configuration.
• Minimum instance = One, for most Figure 4.11 Pooled SOC.exe processes provide
popular map service configurations. optimum execution throughput.
o Provides rapid service response
for first user access. Figure 4.11 identifies ArcGIS Server
o Reserves memory and capacity ArcSOC process pooling options. Process
space for other active map pooling settings are configured when
services. publishing the service.

Pooled ArcSOC processes handle service • The minimum number of instances
requests for multiple concurrent client would normally be “1” to avoid SOC
sessions. startup delays when the service is
requested by a single user. If the service
• Pooled services can be supported by is seldom used, this setting could be “0”,
low or high isolation ArcSOC processes. in which case there would be a SOC
• Pooled services provide optimum server startup delay with the first service
utilization and service throughput. request.
• The maximum number of instances
The pooled service model supports all setting establishes the maximum system
service configurations. The current-state capacity the GIS Server can expose to
information (extent, layer visibility, etc.) is this service configuration. It’s worth
maintained by the Web or client browser repeating: configuring service instances
application. The deployed service instances beyond the maximum platform capacity
are shared with inbound users and released will not increase system throughput, but
after each service transaction for another it may very well reduce user display
client assignment. performance and system stability.
Best Practice: Proper min/max service Minimum and maximum service instance
instance settings are critical to setting:
optimizing service performance and
maximum service throughput. • Establishes parameters for GIS Server
service instance management.
o Minimum service instances are
ArcGIS Server pooling settings (service deployed during GIS Server
instance min/max settings) startup and after idle instance
timeout.
o Maximum service instances
establish limit for GIS Server
service instance deployment.
Timeouts (reference Figure 4.12):
• Service instance (use)

Maximum time a client can use a
service. Establishes timeout for
hung service instances.
• Service queue (wait)
Maximum time a client will wait to
get a service. Establishes maximum
Figure 4.12 ArcGIS Server pooling settings are used service queue time.
to select service configuration minimum and • Service request (idle)
maximum service instance configurations. Maximum time an idle instance
can be kept running. Reduces
Figure 4.12 ArcGIS Service Editor pooled number of deployed service
service instance settings. The service instances to minimum setting when
instance parameters are identified under request for this service has not been
Service Editor Pooling. received for this amount of time.

Low isolation provides more efficient
instance capacity adjustments.
• Single process start-up deploys
ArcGIS Server process configuration multiple service instances.
• Several service instances share the
same process memory.
For example, a GIS Server deploying 12

service instances using high isolation would
be launching 12 separate ArcSOC
processes each providing one service
instance thread. A GIS Server deploying the
same 12 service instances using low
isolation (4 instances per ArcSOC process)
would launch 3 separate ArcSOC
processes, each with four (4) service
Figure 4.13 Server executables can be configured as
either high isolation or low isolation processes. instance threads.
Figure 4.13 shows the available ArcGIS The low-isolation SOC configuration
Server SOC process configurations. There requires less host machine memory; but if
are two types of SOC executable one service instance (thread) fails, the SOC
configurations, high isolation and low process will fail along with its remaining
isolation. Process isolation settings are instances. When a high-isolation service
configured when publishing the service. instance fails, the SOC executable failure is
isolated to loss of a single service instance.
The isolation setting determines how the
server manages the ArcSOC processes. Best Practice: High isolation processes
provide the most stable solution for
• Each SOC process represents a unique standard mapping services.
published map service.
• High isolation deploys single-threaded
ArcSOC processes (one service
instance per process).
• Low isolation deploys multi-threaded ArcGIS Service Editor processes
ArcSOC processes (up to 256 service settings
instances per process).
• Each thread (service instance) is follows
a pointer within the ArcSOC process
tracking execution of the assigned
service request (all requests share the
same copy of the executables).
High isolation provides more stable

operations.
• Each process contains only one
service instance.
• If the service instance fails, only one Figure 4.14 ArcGIS Server Processes settings are
process fails. used to select low or high process isolation. Maximum
number of instances is defined for low isolation
settings.

Figure 4.14 shows the ArcGIS Service Cached map service
Editor processes settings. During the map
publishing process, the author will select the
SOC process isolation, identify the
maximum number of instances per ArcSOC
process (low isolation only), identify recycle
start times, and set the keep alive
connection cycles.
Select SOC isolation.

• High isolation deploys single-threaded
ArcSOC processes (one service
instance per ArcSOC process)
• Low isolation enables multi-threaded
ArcSOC processes (multiple service Figure 4.15 Cached datasets provide multiple scales
instances per ArcSOC process) of tiled images for use as a preprocessed map
Set maximum instances per ArcSOC service. Cache structure includes four times the
number of tiles for each doubling of the scale
process. resolution.
• Low isolation only, can support up to
256 instances per ArcSOC process. ArcGIS Server provides map cache services
• ArcSOC instances are managed by the for use as a data source for Web
GIS Server, based on service applications and ArcGIS Desktop or custom
configuration settings. desktop clients developed with the ArcGIS
Warning: Low isolation GIS Server Engine or ArcGIS Runtime SDKs. ArcGIS
configurations with several instances Server can also stream pre-processed map
per SOC can result in unstable server cache images to ArcGIS Engine and ArcGIS
performance Desktop with 3D Analysis client for 3D
visualization. In both cases, once
Recycle time schedules GIS Server service
downloaded, the cached images are stored
instance recycle operations.
at the client for high performance display.
• Cycles through and sequentially stops
and restarts each ArcSOC instance.
• Schedule during off-peak hours (batch ArcGIS supports on-the-fly translation of 2D
process load sequentially restarting map cache to 3D images with both map
each instance). cache and cached Imagery base maps. For
Keep alive connections. optimum 3D performance, tiles can be
• Checks and repairs data connections for delivered from a 3D Globe cache.
idle instances. Set inspection schedule if
you identify problems maintaining stable Figure 4.15 provides an overview of the
data connections. cached map service image structure.
ArcGIS Server includes automated
geoprocessing services to build and
maintain (pre-process) optimized map
cache pyramids.
Best Practice: Cached tiles enable a

highly scalable static map service.
The cached map service would consist of a

pyramid of pre-processed data imagery or
vector data, starting at a single map

resolution at the highest layer and breaking Draw this map service:
each image into four images at twice the • Dynamically from the data (do not cache
map scale for each additional layer included the source data)
in the pyramid. • Using tiles from a cache (use a cached
tile service)
Tiles are pre-rendered at fixed scales. Caching settings
• Use common ArcGIS Online, Google, • Suggesting tile scheme is provided as
and ArcGIS Online world map scales for default. You can modify default setting
optimum data integration and mash-up to identify minimum and maximum map
performance. scale levels that you want cached.
Rapid display of static base maps. • Estimated cache size is calculated
• Use the map cache to serve static based on the scale selection
basemap layers for the best display Identify caching method
performance. • Build cache automatically
Richer symbols and more information (CachingTools geoprocessing service is
available with high-capacity publishing used to generate cached tiles when
capability. publishing the service)
• Create high-quality basemaps. High- • Build cache manually (Manage Map
quality map cache tiles display as fast Services Cache tools can be used to
as simple tiles. configure your own caching instances to
generate the map cache)
Map service caching configuration Warning: Map cache service instances
will consume CPU resources. Do not use
the ArcGIS Server production site for
manually building your cache
Best Practice: Use preconfigured
CachingTool instances when generating
map cache. Batch processing loads for
maximum CachingTool service instance
configuration must be included in your
capacity planning analysis to ensure
adequate server resources during peak
loads.
Figure 4.16 Service editor caching capabilities allow

you to specify if the service will be cached during the
map service publishing process.
Figure 4.16 shows a view of the ArcGIS

Service Editor Caching settings. A Portland
dataset is selected and the display shows
the suggested tiling scheme. A tool is
included that calculates an estimate of the
cache size. You can choose to publish a
dynamic or cached service. You can have
the cache creation start automatically once
you complete the configuration, or save the
configuration and start the caching job
manually at a later time.

Advanced settings • Generate cache on the fly. (This will
generate a map cache for an area
including the display extent with the first
service request and add the tiles to your
map cache. Subsequent service
requests for the same extent will use the
cached map tiles). Published map
service instances are used to generate
the on-demand cached tiles while
servicing the map request.
Best Practice: Cache on demand should

only be used for generating cache on
sparsely populated areas (map request
Figure 4.17 Service Editor advanced service caching for that area of the map would be a rare
parameters identify when and how the service event).
caching job will be run.
Additional Advanced caching
Figure 4.17 shows three options available settings
for map service caching.
The Service Editor Caching Advanced
• Generate full cache of available data Settings tab includes an additional
set. (This will generate a map cache Advanced Cache Settings tab (Figure 4.14)
based on your caching specifications for that is important to review. This tab includes
the entire dataset based on your three very important performance settings:
minimum and maximum scale
selections)
• Storage Format setting identifies how
the cached tiles will be stored. The
Warning: Map cache processing COMPACT setting stores the cached
instances will consume CachingTools tiles in bundles for reduced storage
service instances while generating the volume and improved file migration
cache. If all published CachingTools performance. A single bundle can store
instances are busy, caching job will wait up to 16,384 (128 x 128) tiles.
in queue until service instances are • Allow clients to cache tiles locally. This
available. setting is a very important performance
consideration for remote clients with
• Generate partial cache of available data limited bandwidth. An initial map display
set. (This allows you to select an area of can require over 10 times the traffic to
interest for caching. This can be defined download cached tiles over the network.
by a polygon layer or buffer that shows Once delivered to the client, the tiles
the areas on the map you wish to can be stored in browser cache and
cache.) local tiles can be used for subsequent
displays. If you do not allow clients to
Best Practice: Partial cache can reduce cache tiles locally, all tiles will be
overall caching time. For most downloaded from the server for each
implementations, service requests display contributing to network
access a relatively small area of the total saturation and poor display
dataset. performance.

Best Practice: Map cache clients should number of tiles in the map cache and the
periodically clear their local browser average map service time for rendering
cache to ensure use of the latest tile each map tile image. Figure 4.18 can be
updates available on the server. used to get a rough estimate of the
expected map cache generate time.
• Allow clients to export cache tiles. This
setting allows clients to export tiles to Map cache generation time is a function of
their local machine for low bandwidth or the number of tiles and average tile
mobile operations (use when not generation time.
connected to the server). Exporting tiles
to the local machine prior to working in • Chart shows one tile at the top layer
the field can significantly improve (multiply result by number of tiles in top
display performance over low bandwidth layer)
connections. • The number of tiles increases by a
factor of four with each additional layer.
Generating the map cache • Tile render time can vary from less than
one second to several seconds,
depending on the average tile map
ArcGIS Server creates cache tiles using a
complexity.
geoprocessing service named
• Processing hours based on single
CachingTools. This service is configured for
CachingTools service instance (divide
you in the System folder when you create
result by total concurrent service
the ArcGIS Server site. The number of
instances used when generating the
instances you allow for the CachingTools
cache).
service determines how much power your
machine can dedicate toward caching jobs.
Rendering time increases exponentially with
each tile layer. With average 0.5 tile render
Estimated map cache generation time time, we expect the following generation
times:
• One (1) hour for seven layers

• Five (5) hours for eight layers
Estimating map cache generation time.
• Build a small area to test the output

symbology, labeling, and performance
criteria for your primary map client.
• Execute cache jobs in sections to
manage production time.
Figure 4.18 Engineering chart for estimating cache
generation time. Processing time grows exponentially Warning: The caching process can be
with each additional layer of cache tile generation. time consuming.
Client access to the cached data would

deliver tiles that correspond to the
requested map scale. Tiles would be
blended together as a single reference layer
by the client application. Total pre-
processing time would depend on the total

Cache processing profile
If you have the service instances configured

properly, you should be able to take full
advantage of the available hardware
processing capabilities. CPU usage should
peak at 100 percent when running with N+1
service instances with hyperthreading
disabled.
Figure 4.19 Caching timelines are reduced linearly
based on the number of available processor core. Warning: ArcGIS Server GIS Servers are
cluster-aware, and common service
Figure 4.19 provides an example of taking instance configurations apply for each
advantage of the hardware, as described GIS server machine within a site cluster.
above. ArcGIS Server will use the If you establish a maximum service
CachingTools maximum available service configuration of 5 instances, then 5
instances to process the map cache as service instances will be deployed for
quickly as possible with the available each GIS Server machine within the Site.
hardware.
This video on [ArcGIS 10.2.1 Map Caching
Cache processing timelines can be in ArcGIS Server] provides a demo showing
improved by increasing the number of the new workflows and features of map
concurrent processor cores utilized in the caching since the ArcGIS 10.2.1 for Server
caching process. Recommended release.
CachingTools service instance configuration
is N+1, where N=number of physical server
core processors. If you want to view caching Manage Services caching tools
status and manage cache jobs while
executing the cache, maximum number of
instances = physical server process core
leaves some processing resources for
managing caching services.
Testing was performed on identical 4-core

server platforms.
• Single service instance (thread) on a

single 4-core server took 500 hours.
• Five (5) service instances (threads) on a Figure 4.20 Managed Services caching tool
4-core server took 125 hours (four times services are configured by the ArcGIS
faster). Server Manager during the initial site install.
• 10 service instances on two 4-core
servers took 65 hours (eight times Manage services caching instances are
faster). configured during site installation to enable
background caching services.
Best Practice: Take advantage of
available hardware resources • CachingControllers: Establish number of
concurrent caching jobs per server

machine supported by the CachingTools Figure 4.21 shows the CachingTools
instances. processes configuration. An identical tab will
• CachingTools: Establish number of be configured for the CachingControllers
caching service instances per server processes configuration.
machine available for caching job
processing. CachingControllers Pooling settings
• You should always run the
CachingTools service and the • Min service instances = 0. There is no
CachingControllers service in the same need to have instances running if there
cluster. are no assigned caching jobs.
CachingController instance will be
Service Editor service configurations apply running to manage each active caching
to every GIS Server machine within the job.
assigned cluster. • Max service instance setting identifies
the max number of concurrent caching
Best Practice: Managed Services jobs. If all available instances are busy,
caching tools should be configured in a the job will be held in the process queue
separate ArcGIS Server Site cluster from until a CachingControllers instance is
production Web services. available.
• Timeouts should match the
Production servers within the GIS Server CachingTools settings below.
Site can be reassigned to the Map Caching
cluster to expand caching capacity during CachingTools Pooling settings
off-peak hours.
Figure 4.21 shows the Pooling settings for
the CachingTools service. The
System Caching Pooling configurations recommended service instance setting for
caching services is N+1 (number of
available core on the server machine plus
A dedicated CachingControllers instance is
one). Maximum setting of 3 service
required for each caching job. Pooling and
instances would be appropriate for 2-core
Processes settings must be configured for
GIS Server machines.
the CachingControllers service.
CachingTools will be configured to optimize
utilization of available platform hardware The CachingControllers manage caching
resources during peak cache processing jobs to take advantage of all available
loads. CachingTool service instances. For a single
caching job, the CachingController will
assign multiple caching processes to
leverage all available CachingTool service
instances (up to the total number of bundles
in the caching configuration). Multiple
concurrent jobs will share available
CachingTool service instances. Assignment
of multiple concurrent caching jobs (multiple
CachingControllers) will optimize utilization
of available server processing resources.
Figure 4.21 CachingTools service instance
configuration should be set based on total GIS server
machine cores (N+1) for optimum caching throughput. Timeout settings should be appropriate to
the caching service.

• Maximum time a client can use a days at a time often generating hundreds of
service = 360000 seconds (provide thousands of cached map tiles. It is
sufficient time to complete the caching important to support this type of service with
job). CachingController process will be the most stable process configuration. For
dedicated for the complete caching job. this reason, both the CachingController and
CachingTools instances will be CachingTools SOC processes should be
reassigned after completing each configured in high isolation mode.
bundled cache assignment.
• Maximum time a client can wait to use a Recycling the processes can ensure stable
service = The caching service is the instances during the map cache build,
client within this configuration, and the promoting optimum performance during
caching service wait time is hard coded cache generation. Caching jobs are
to around 30 days (use default setting). processed in bundles of 16,384 tiles.
• Maximum time an idle instance can be CachingTools instances can be recycled
kept running = 180 seconds (this can be between processing bundles once during
set for a short time, allowing instances each recycle process. CachingControllers
to be shut down if not in use). and CachingTools processes configurations
should be scheduled for a recycle every 24
Best Practice: Think carefully what hours to promote optimum site stability and
works best for your Site's caching jobs performance. CachingController instances
and select appropriate service instance will recycle between caching jobs.
configurations and timeout settings to CachingTools instances can recycle during
optimize utilization of your server job processing between bundle caching
environment. assignments.
System Caching Processes

configurations GIS Server machine
memory configuration
Figure 4.22 CachingControllers process configuration

should be set at high isolation for optimum stability.
Figure 4.20 Configuring the server machine with
sufficient physical memory is important for ensuring
Figure 4.22 shows the CachingController performance and capacity goals are met.
processes configuration. An identical tab will
be configured for the CachingTools The platform configuration must include
processes configuration. sufficient physical memory to accommodate
all active process executions. Systems that
Map cache generation is an intense batch do not have sufficient memory will
process which can often run for hours and experience performance degradation during

peak load conditions (processing overhead • Minimum of three instances per core is
increases due to operating system reasonable.
swapping executables between disk and
memory). If memory is not sufficient to Best Practice: Provide sufficient memory
support concurrent server processing to support optimum performance.
requirements, system will start to
experience random process failures ArcGIS Server memory
(process executables dropped from memory recommendations
during execution). Figure 4.20 shows
memory performance considerations for an
ArcGIS platform memory recommendations
ArcGIS Server host machine deployment.
• Minimum of 4 GB memory per core
Best Practice: Sufficient platform
recommended.
physical memory is critical to successful
• Additional 1 GB memory per host
stable operations. More memory can
platform core recommended for virtual
improve performance.
server host platforms.
• More memory may be required when
Active software processes must be in using large file data sources (imagery).
physical memory for successful execution. If • More memory may improve throughput
sufficient memory is not available: performance.
• Inactive processes will be swapped to The operating system will use extra memory
memory cache to make room for active for data cache, which for some workflows
processes. can improve throughput performance.
• Active processes will crash if swapped
from memory during execution.
Managing host platform service
Too many instances per server can exhaust
memory.
instance capacity
• Increased paging when not enough There are a handful of GIS Server terms
memory. and configuration variables – there is no
• Slower processing due to shared simple recipe for getting it right. It all
compute resources. depends on your user environment – how
popular will your services be and how will
people access your site. The overall goal is
Optimum total active service instance
to configure GIS Server to handle the
assignments can vary based on service
maximum number of requests with the
popularity. Limiting maximum of 10
minimum amount of processing overhead. A
instances per core is reasonable for most
summary of the service configuration and
map service scenarios.
optimum capacity discussion for ArcGIS
10.0 for Server, along with a nice summary
• Provides sufficient overhead for server diagram, is available in Chapter 4: Software
instance management. Performance 30th Edition.
• Avoids excessive starting and stopping
ArcSOC processes during peak
ArcGIS Server manages the internal
throughput loads.
ArcSOC service instance deployment based
on instructions provided with each service
Too few instances per server: configuration. Services are what you publish
on the GIS Server. Web applications
• Can limit utilization of host hardware. consume these services to produce the

client display. The service requests are sent instances required for peak system
to ArcSOC process threads (service throughput.
instances) that are executed by the
hardware core processors. During installation, GIS Server is installed
on each host machine. This provides the
Each published service will have a service ArcGIS code needed to support SOC
description file. The service description file deployment. During startup, the GIS Server
will contain the parameters you identify deploys the minimum number of service
when publishing each service description. instances for each service description within
How you configure your services will each GIS Server machine. Service
determine how they will be deployed by the instances are deployed in SOC.exe
GIS Server. Service description files are processes.
created from ArcGIS Desktop (ArcMap)
during the map publishing process. During operations, if concurrent requests for
a specific service exceed the number of
There are some key things to keep in mind deployed service instances, the GIS Server
when publishing services. The GIS Server will increase the number of deployed
will deploy SOC service instances within the service instances to handle peak request
bounds established by the service rates up to the maximum value allowed for
description MIN/MAX instance settings. The in the service description. If the inbound
minimum service instances for all services concurrent requests for that service exceed
will be deployed during GIS Server startup – the maximum deployed instances you’ve set
this establishes the minimum number of in the service description, requests will wait
instances that will be deployed within each in the service queue until an existing service
GIS Server machine. During peak instance is available for service assignment.
concurrent loads, the GIS Server can Non-active services can be reduced down
increase the number of SOC instances up to the minimum instances specified in their
to the maximum service description levels to service description file based on their
support concurrent inbound service service instance timeout settings (max idle
requests. time).
Keep in mind, there is extra platform Deployment algorithms within the GIS
processing overhead required every time Server provide even distribution of service
the GIS Server has to start a new SOC instances across the assigned host
process. Ideally, you would like to deploy platforms. The deployment algorithms along
just the right amount of service instances so with the service queue work to balance the
there is one available for immediate GIS GIS Server processing load across
Server assignment for each client request. machines assigned to the same GIS Server
During peak server loads, you want to have Site. The GIS Server will be used as the
just the right number of maximum service final load balance solution for the machines
instances identified to fully utilize available within the GIS Server site.
host platform compute resources staying
well within available platform memory Understanding what each of these
limitations. During maximum peak loads, performance parameters do and how they
you want to limit concurrent processing should be configured to satisfy your specific
loads to allow optimum service throughput. service needs is important for optimum
For high performance services (processing utilization of your host servers. Getting this
time less than 1 second) you may want to right will take some thinking and some
allow the total of several active services to careful planning. Once you deploy your
be two or three times the number of services, you will need to monitor the

service logs to see if your settings are Greek citizen case study background
working. Modifications can be applied to
optimize your specific service loads. A
sufficient number of service instances must
be deployed to consume available host
platform compute resources and enable the
host platform to service peak throughput
loads. Sufficient memory must be available
to handle the maximum number of active
service instances or the system can
become unstable. Remember, the goal is to
configure the system to minimize
processing overhead during peak
throughput periods (excessive SOC process
start-up loads during peak service periods is
overhead you would like to avoid).
Additional ArcGIS platform memory Figure 4.21 Greek government contacted Esri to
configuration guidelines are provided in the develop a solution for their Greek National Citizen
appendix on Windows Memory Declaration.
Management.
The customer had a requirement to design
a web application solution that would be
used to collect national property location
Selecting the right and census information during a three-
technology: A case study month national citizen declaration period.
The Figure 4.21 shows the area of Greece
Selecting the right software technology can involved in the census.
make a big difference in performance and
scalability, and cost of the production Citizens would report to local regional
system. The following case study shares an government centers and use a local desktop
experience with a real customer computer to locate their home residence on
implementation which clearly represents the a map display generated from a national
value of selecting the right software imagery and geospatial feature repository.
technology. The citizen would place a point on the map
identifying their residence, and then fill out a
reference table identifying their census
information.
The citizen input would be consolidated at a

centralized national data center and shared
with all regional government centers
throughout the declaration process.

User requirements for web mapping Web mapping services architecture
solution patterns.
Figure 4.22 This network diagram shows the central

National Data Center and sample small and large Figure 4.23 Three ArcGIS Server web technology
regional centers used in completing the design. patterns were considered. They included an early
Map Editor ADF application, two Adobe Flex
Figure 4.22 provides an overview of the applications, and one Windows Mobile application.
national architecture. The initial system
design was developed using ArcGIS Server Figure 4.23 shows the ArcGIS Server
feature service with an imagery tile cache architecture patterns that were considered
with browser clients located at 60 regional for the Greek citizen declaration solution.
national sites. Following contract award, the
customer reviewed available technology Initial hardware proposal
options to finalize the system design.
The following workflow was used to
Peak web service use requirements generate system loads for the initial
hardware proposal.
• 2400 concurrent client edit sessions.
o 75 percent map query to find • Web REST map services with central
home location dynamic enterprise geodatabase data
o 25 percent simple edits (select layers and imagery tile cache.
point and complete attribute CPT Workflow: AGS REST 2D V Lite
table) 100%Dyn 13x7 Feature +$$
• 60 remote user locations, one central System implementation design
national data center. review (after grant approval)
o Large site: 50 concurrent clients
o Small site: 10 concurrent clients After some time, the European Union
approved the Greek Citizen Declaration
grant based on the initial hardware
proposal. The Greek cadastral team
traveled to Esri to review available
technology options for final implementation.
The following web mapping services

architecture patterns were reviewed to
identify optimum deployment scenario.

• Web REST map feature service with Web map single point feature service
central dynamic SDE data layers and layer with central map cache
Imagery cache
CPT workflow: AGS REST 2D V Lite CPT Calculator Analysis shows a Web map
100%Dyn 13x7 Feature +$$ single point feature service layer with
• Web map service with point feature central map cache.
layer + central map cache
CPT Workflow: AGS REST 2D V Lite ArcGIS Server provides a data cache option
10%Dyn 13x7 Feature +$$ where reference map layers could be pre-
• Web mobile application + local map processed and stored in a map cache
cache pyramid file data source. Pre-processing the
CPT Workflow: AGS SOAP 2D V Lite reference layers would significantly reduce
5%Dyn 13x7 Feature server processing loads during production
Best Practice: Significant technology operations. A single point declaration layer
improvements have become available contained all features that would be edited
since the initial proposal. It is always and exchanged during the citizen
good to update the final solution declaration period; all remaining reference
architecture based on current layers could be cached. Changes would be
technology before final implementation. displayed at all remote site locations with
each client display refresh.
Web map feature service with central
dynamic SDE data layers and imagery Hardware solution:
tile cache • 2 Xeon E5-2637v4 4 core (1 chip)
3500 MHz servers
CPT Calculator analysis shows platform • ArcGIS Server licensing for up to 4
solution when supporting business needs cores
using 100 percent dynamic web map Total estimated technology cost
feature service and remote client browser • Hardware cost = $18,474
technology. • Software licensing = $12,000/year
Peak network traffic estimates
• 5.9 Mbps for large sites, recommend
Hardware solution:
12 Mbps bandwidth
• 4 Xeon E5-2637v4 4 core (1 chip) 3500
• 1.2 Mbps for small sites, recommend
MHz servers
3 Mbps bandwidth
• ArcGIS Server licensing for up to 8
• Estimated network cost of
cores
$1,901/month.
Total estimated technology cost
• Hardware cost = $36,102
• Software licensing = $24,000/year Web mobile application with edit feature
Peak network traffic estimates synchronization + local map cache
• 36 Mbps for large sites, recommend 90
Mbps bandwidth CPT Calculator analysis. The third design
• 7.2 Mbps for small sites, recommend 18 option was to use the ArcGIS Mobile
Mbps bandwidth application with a local reference cache
• Estimated network cost of data source. A demo of the ArcGIS Mobile
$8,177/month. client was provided on a Windows desktop
platform to demonstrate feasibility of
supporting the required editing functions
with this client technology. The ArcGIS
Mobile client technology operates very well
on a standard Windows display environment

and performed all the functions needed to network traffic and improve display
support the citizen declaration performance even more. Selecting the right
requirements. technology can make a big difference in
total system cost and user productivity.
Hardware solution: Figure 4.24 highlights the advantage of
• 1 E5-2637v4 4-core (1 chip) 3500 selecting the right technical solution.
MHz servers
• ArcGIS Server licensing for up to 4
cores Best Practice: Selecting the right
Total estimated technology cost: technology solution can make a big
• Hardware cost = $8,814 difference in price and user
• Software licensing = $12,000/year performance.
Peak network traffic estimates:
• 1.7 Mbps for large sites, recommend Business analysis identifies clear
3 Mbps bandwidth advantages in the ArcGIS Mobile solution.
• 0.33 Mbps for small sites,
recommend 1.5 Mbps bandwidth • Over $27,288 savings in hardware cost
• Estimated network cost of alone.
$563/month • Significant additional savings on
software licensing and reduced
infrastructure network bandwidth costs.
Caching advantage summary
Server Software
performance summary
Experience suggests we can do a better job
configuring ArcGIS Server deployments.
Figure 4.24 Business analysis comparing the four Proper server configurations can reduce
software technology solutions.
implementation risk and save customer time
and money. Projects can be delivered within
It was very clear that the cached client project cost, time, and satisfy performance
application provided significant cost and budgets.
performance benefits over the centralized
Web application dynamic solution included
in the initial proposal. Pre-processing of
map reference layers as an optimized map CPT Capacity Planning videos
cache pyramid can significantly improve
display performance. Use of an intelligent The next section will take a closer look GIS
desktop client that can access reference Data Administration.
layers from a local map cache can minimize
Page Footer
System Design Strategies 26th edition - An Esri ® Technical Reference Document • 2009 (final
PDF release)

GIS Data Administration
Fall 2017 GIS Data Administration 41st Edition
Data provides the resources you need to GIS feature data

make proper business decisions. The
information products required to make architecture
business decisions determine the critical
data resources that must be available for
business operations. How you organize and
maintain your data resources will contribute
to system performance and user
productivity.
How GIS data is managed has changed

dramatically over the past 10 years. Much of
this change is driven by technology. The big
focus in the 1990s was to move GIS data
resources together in an enterprise
geodatabase, where users could better
manage and share enterprise data Figure 5.1 GIS feature data architecture includes
resources. Data management today Enterprise geodatabase (production source),
geodatabase (GDB) archive, distribution database
includes multiple publication formats to (publication data source), and vector map cache.
improve display performance and capture
change over time. Figure 5.1 shows the data source
architecture patterns available to manage
A variety of data management and GIS feature data. GIS feature data includes
distribution strategies are available today to points, polygons, lines, complex features,
improve data access and dissemination and associated attributes. Additional content
throughout the rapidly expanding GIS user may include parcel fabric, cartographic
community. The volume of data you must representations, lidar point elevation, terrain
sort through each day is growing data, etc.
exponentially. How you manage, organize,
and control these data resources is critical Geospatial data is the core integration of
to your success. business intelligence. How you organize
your data contributes directly to your
business complexity and drives the
performance of your business operations.
Good data management empowers your
ability to make proper business decisions.
There are only two kinds of data: useful

data and useless data. Useful data is what
you use to create business information
products and enable informed business
decisions. Useless data is what you do not
use, and can rapidly increase the
complexity of your data repository.
GIS Data Administration 5.1

Best practice: Data should be organized Best practice: Distribute geospatial
and managed to empower proper business operational layers in a
business decisions and optimize user publication geodatabase.
productivity.
GIS Feature Data Map Cache
GIS Feature Data Production
Database A feature data map cache is a collection of
preprocessed tiled map images stored at
A production data source is an Enterprise multiple scales for rapid dissemination.
geodatabase used to organize and manage
your geospatial feature data resources. • Combines multiple geospatial layers into
multiple levels of read-only
• Maintains a complete collection of all preprocessed map tiles.
critical geospatial feature data. • Configured in a pyramid tile structure at
• Includes schema for managing and standardized projection and map scales.
validating accuracy and integrity of • Cached format delivers map tiles with
spatial feature edits. negligible processing overhead.
• Includes functional dependencies and • Structured tile format enables client
relationships between feature datasets. browser caching for high display
• Manages multi-user versioned edit performance.
operations. • Optimum distribution format for static
• Includes all work in progress (versions) basemap layers.
along with published datasets
(DEFAULT). Best practice: Distribute geospatial
static basemap layers in a preprocessed
Best practice: A production data source map cache.
provides a single integrated repository
for all enterprise-level geospatial feature GIS Feature Data Archiving
data resources.
Geodatabase (GDB) archiving (shown as an
GIS Feature Data Publication available component of the production data
Database source) includes functionality to record and
access changes made to all or a subset of
A publication geodatabase is an Enterprise data in a versioned geodatabase.
or file geodatabase used to optimize
distribution of finalized geospatial data • Provide a mechanism for capturing,
resources. managing, and analyzing data change.
• Creates and maintains a separate
• Includes read-only simple feature feature class schema associated with
geodatabase. the versioned geodatabase.
• Provides simple feature format that • When enabled, maintains all changes
improves display performance and saved or posted to the DEFAULT
system capacity. version in an associated archive class.
• Provides separate distribution access • Enables temporal analysis of geospatial
layer improving data security. resources over time.
• Provides optimum distribution format for
operational geospatial layers.

ArcGIS Data Stores named versions are proposed changes and
database updates in work.
The Esri managed data repository that
stores spatial content that has been shared What is versioning?
to Portal.
• Relational Data Store manages feature

content published through Portal.
• Tile Cache Data Store manages 3D
Scenes and raster layer content
published through Portal.
• Spatiotemporal Data Store manages
temporal feature content for GeoEvent
Server and GeoAnalytics Server output
layers.
Figure 5.2 A versioned geodatabase manages

multiple edit sessions over time without duplicating
data.
CPT Platform Capacity Calculator Custom
Web Mapping Services Geodatabase versioning allows multiple
users to edit the same data in an Enterprise
The CPT Platform Capacity Calculator is a geodatabase without applying locks or
simple tool for evaluating selected platform duplicating data. Figure 5.2 provides a
capacity. The default tool, located at the drawing of a versioned geodatabase
bottom of the CPT Hardware tab, includes a workflow.
variety of standard workflows that
demonstrate platform capacity. For analysis Users always access an Enterprise
and reporting purposes, you may want to geodatabase through a version. When you
change the default list of sample workflows connect to a multiuser geodatabase, you
and include those workflows you are specify the version to which you will
evaluating in your own design environment. connect. By default, you connect to the
This link describes how you can change the DEFAULT version.
Platform Capacity Calculator workflow
samples to a custom set of workflows for Best practice: Use a versioned
demonstration purposes. geodatabase when managing multiple
edit sessions of common feature
Enterprise Geodatabase datasets over time.
Release of ArcGIS technology introduced Geodatabase versioning example

the Enterprise geodatabase, which provides
a way to manage long transaction edit
sessions within a single database instance.
Enterprise supports long transactions using
versions (different views) of the database. A
geodatabase can support thousands of
Figure 5.3 Typical long transaction workflow lifecycle
concurrent versions of the data within a includes an initial edit session to develop a prototype
single database instance. The default design, a relatively long construction phase, and a
version represents the real world, and other final as-built design update before final posting.

GIS users have many use-cases in which an additional state to the version
long transaction workflows are critical. lineage.
Figure 5.3 shows a long transaction • The edit version represents the Edit
workflow for developing a new community state lineage.
housing subdivision. • A new user edit session is started from
DEFAULT lineage state 1.
A new housing subdivision is being • User edit session saves state 1a and 1b
approved by the city. when completing new housing
subdivision design.
• City submits requests for design • Other user edit sessions are posted to
proposals for the new subdivision. DEFAULT, creating states 2, 3, and 4.
• City planning establishes edit sessions • User edit session completes the as-built
and develops multiple subdivision design and begins to reconcile process.
proposals. • Any row conflicts (deletions, additions,
• Subdivision design proposals are or modifications by other user edit
provided to the city council for review sessions) are identified during the
and approval. reconcile process.
• Design is selected and approved for • Once conflicts are resolved, the edit
construction. session can be posted to DEFAULT,
• New housing subdivision is constructed creating new state 6.
over a two-month period.
• City planning updates design for as-built ArcSDE version state tuning
subdivision.
• New housing subdivision is posted to
DEFAULT for publishing and
distribution.
ArcSDE explicit state model
Figure 5.5 ArcSDE version state tuning functions

provide administrators with tools to improve database
Figure 5.4 Versioned geodatabase includes a performance.
DEFAULT lineage and multiple open version lineages
saved at various states over time. Figure 5.5 shows the ArcSDE Geodatabase
DEFAULT version state tree. Enterprise
Figure 5.4 shows the progress of a production geodatabase maintenance
versioned workflow edit session over time. environments often support many GIS
editors posting many changes to the
The diagram shows DEFAULT version geodatabase DEFAULT lineage over time.
lineage and new version lineages. In many scenarios, the DEFAULT lineage
tree can rapidly grow to hundreds and even
• The DEFAULT version is the current thousands of state changes. Many of the
"public" end-user view of the states are redundant and can cause
geodatabase. database tables to grow. Reducing the size
• Lineage refers to the version states as of these tables can improve database
they are updated. Each update provides performance and reduce maintenance
overhead.

Compressing the state tree removes each Versioned geodatabase view
state that is not currently referenced by a
version and is not the parent of multiple
child states. When executed by the ArcSDE
administrator, all states that meet the
compression criteria are removed,
regardless of owner. All other users can
compress only the states that they own.
This operation reduces the depth of the
state tree, shortening the lineage and
improving query performance.
Best practice: DEFAULT state tree

should be compressed on a periodic
schedule to maintain optimum query
performance. Figure 5.6 Several tables are added to the
geodatabase schema when establishing a versioned
geodatabase. Two key tables include the Adds table
You can also trim the state tree, which and the Deletes table.
collapses a linear branch of the tree. This is
done by reconciling and saving an edit
Figure 5.6 shows the key database tables
lineage to a more current state (for
supporting a versioned geodatabase view.
example, version t1 could be reconciled and
Several additional tables are included in a
saved to reference state 7, freeing state 1
versioned geodatabase. ArcSDE uses the
for compression). Trimming database states
additional tables to manage multiple
will reduce the depth of the state tree,
concurrent edit sessions and query access
shortening the lineage and improving query
to different views of the data.
performance.
For example, a user query of the current
Best practice: The DEFAULT state tree
DEFAULT view would include features
may need to be trimmed to reduce
(rows) in the Base table, plus any posted
number of long transaction reference
rows from the Adds table, minus any posted
states for optimum query performance.
rows from the Deletes table. This same
approach would be used to view open Edit
versions, with ArcSDE sorting out what
features were required to populate the
version view.

Versioning managed by ArcSDE of the geodatabase. ArcSDE also supports
schema export and import of data from and to the
appropriate database tables and maintains
the geodatabase scheme defining
relationships and dependencies between
the various tables.
Figure 5.7 Enterprise geodatabase includes the

ArcSDE schema and the user schema.
Figure 5.7 shows the Enterprise

geodatabase schema. The Enterprise Figure 5.8 There are four basic types of geodatabase
geodatabase includes the ArcSDE schema replication patterns that enable a broad variety of
and the user schema. The Enterprise distributed enterprise and federated architecture
geodatabase license key must be installed scenarios for GIS deployment.
with the ArcSDE schema.
ArcSDE provides a variety of data
The ArcSDE schema is used to manage replication options associated with a
operations within the versioned database versioned geodatabase as shown in Figure
environment. 5.8.
• Query views to DEFAULT or to specified Note: Replication is the process of sharing

open edit versions of the database. data to ensure consistency between
• Reconcile and post operations related to redundant sources. Geodatabase
the various versioned states. replication provides filtered data sharing at
• Geodatabase archive schema for the version level.
managing state history.
• Geodatabase replication services for Mobile operations:
managing versioned updates sent to
distributed geodatabase sources. • Check-out and synchronization with
• Geodatabase licensing to protect mobile laptop clients
geodatabase functional integrity. • Provisioning and synchronization with
ArcGIS Windows Mobile clients
Note: Multiple user schema instances were
supported starting with the ArcGIS 9.2 Production/publication:
release.
• Incremental updates to an Enterprise
Geodatabase replication geodatabase publishing database
• Incremental updates to a file
use-cases geodatabase publishing environment
ArcSDE manages the versioning schema of

the geodatabase and supports client
application access to the appropriate views

Distributed operations: ArcSDE geodatabase replication support
for disconnected editing operations.
• Corporate production Enterprise
geodatabase synchronized with multiple
remote office production Enterprise
geodatabases
• Centralized scalable ArcSDE distributed
geodatabase architecture sharing a
common storage area network
Figure 5.9 Mobile desktop operations support ArcGIS
Hierarchical operations: client remote field editing.
• Local production geodatabase sharing Figure 5.9 shows ArcSDE geodatabase

versions of their data for regional and replication support for disconnected desktop
national operations SDE geodatabase client editing operations.
• Global remote operations exchanging
versions of their data with regional and Note: Disconnected editing extends the
corporate management operations geodatabase to provide clients with the
capability to perform edit operations in the
Mobile operations field when not connected to the central
geodatabase.
Mobile field operations are a big part of GIS
workflows within most organizations. Check-out operations were initially
supported with the ArcGIS 8.3 release.
• Many mobile operators do not have
direct input to the computerized • ArcGIS editor opens edit session with
business workflow processes. production geodatabase (creates edit
• Often information collected in the field version).
must be entered into the system once • Check-out of work area operational
the users return to the office. layers to desktop SDE geodatabase
• In many cases, there is a staff of editors (e.g., SQL Express).
who enter material from marked-up • Check-out reference layers to mobile file
hardcopy provided by the field geodatabase.
operators. • Complete disconnected field edit
operations.
Best practice: Mobile operations provide • Check-in field edits to version on central
a way to integrate field operators into the production geodatabase on return.
computerized business workflow
processes, improving business
efficiency and reducing effort required to
get data into the computerized systems.

ArcSDE geodatabase replication support • Default version queries
for disconnected Workgroup SDE • Multiple edit version sessions
geodatabase editing operations. • Reconcile and post operations
• Data schema table dependencies and
relationships to maintain data
consistency
• Geodatabase history archiving
• General maintenance operations
For many GIS operations, hundreds of

Figure 5.10 Mobile workgroup operations support users throughout the organization require
remote SDE geodatabase field operations with
multiple editors.
access to the production data source, most
requiring access to the published DEFAULT
version.
Figure 5.10 shows ArcSDE geodatabase
replication support for disconnected
Workgroup SDE geodatabase server editing Best practice: Separating the publishing
operations. database from the production database
provides a more scalable and secure
data management environment.
Check-out operations were initially
Several reasons an organization may want
to use a separate publication geodatabase:
• ArcGIS Editor opens an edit session
with the production geodatabase
(creates edit version). • More scalable server architecture
• Check-out of work area operational (distributed database loads).
layers to workgroup SDE geodatabase • More secure production environment
(e.g., SQL Express). (viewer access limited to a publication
• Check-out reference layers to mobile file data source).
geodatabase. • Expand data center capacity
• Deploy for disconnected field edit (publication database can be hosted by
operations. cloud vendor).
• Multiple editors can check out from • Limit public access to DMZ (publication
workgroup SDE geodatabase for mobile database can be located in the DMZ).
field operations.
• Reconcile and post field edits in mobile
workgroup SDE geodatabase.
• Check-in field edits to version on central
production geodatabase on return.
Distributed Geodatabase Operations can be

used to support disconnected SDE
geodatabase clients with incremental
synchronization capabilities.
Figure 5.11 One-way replication from a production
geodatabase to a publication geodatabase.
Production/publication operations
Figure 5.11 shows how ArcSDE replication
A versioned production geodatabase is used to share a version of the Enterprise
performs many functions related to the geodatabase on a separate publication
editing workflows that place processing geodatabase.
demands on the server.

• Feature-level check-out to production • Provides web access to a separate copy
SDE geodatabase was supported with of the published dataset.
the ArcGIS 8.3 release. • Filtered versions of the production
• One-way incremental multi-generation geodatabase can be distributed to
checkout to SDE geodatabase was separate publication instances, based
supported with the ArcGIS 9.2 release. on enhanced security requirements.
• One-way incremental multi-generation
checkout to file geodatabase was Extract/transform/load operations
Best practice: ArcSDE replication is the

best solution when moving part of an
Enterprise geodatabase to a separate
geodatabase instance.
There are several advantages to using a

separate publication geodatabase instance
Figure 5.12 Geodatabase transformation between
for sharing data to GIS viewers. geodatabases with different schema is provided by
the data interoperability extension.
Organizations use one-way geodatabase
replication for the following reasons: Figure 5.12 shows how ArcSDE
geodatabase replication can be used to
• Improved performance and system move data to a different geodatabase
scalability. schema. Geodatabase transformation is a
o Replicating to a simple feature term used to represent replication to a
(DEFAULT) read-only geodatabase with a different schema.
publication database can
improve query performance and Best practice: Geodatabase transform is
increase platform capacity. the best solution for replicating part of a
geodatabase to a separate geodatabase
Best practice: An SDE geodatabase instance with a different schema.
should be used when live updates are
required during peak viewing operations. The ArcGIS Desktop Data Interoperability
extension is used to create a script to
• Replicating to a file geodatabase can transform data between the two schema.
reduce DBMS processing loads and
improve display performance. Best practice: A service can be deployed
on ArcGIS Server for incremental one-
Best practice: A separate file way replication.
geodatabase instance should be
provided for local access by each GIS
server for optimum query performance.
Separating the publication database from

the maintenance database improves data
security.
• Limits direct user access to the

production geodatabase.

Distributed geodatabase operations located on the parent corporate
geodatabase.
Distributed remote ArcGIS editor

(disconnected geodatabase editing
operations):
• Version replica provided to ArcGIS

editor desktop geodatabase (SQL
Express).
• Reference data layers are replicated to
local file geodatabase (one-way
geodatabase replication).
• ArcGIS editor can work in a
Figure 5.13 Distributed geodatabase replication disconnected mode when mobile.
enables functionality for a single SDE geodatabase • ArcGIS editor synchronizes changes
schema synchronized across multiple SDE
geodatabase platforms.
with corporate when connected.
Figure 5.13 shows a distributed Best practice: Distributed geodatabase

geodatabase configuration. ArcGIS replication is convenient for single-user
Geodata services can be used to establish mobile editing operations.
distributed geodatabase operations for
scale-out database architectures. Geodatabase replication can be used to
build and support central data center high-
Distributed regional office support: capacity scale-out geodatabase operations:
• Centralized corporate geodatabase • Parent production geodatabase

represents the parent production established for regional production data
database. integration.
• Versioned replica is provided to • Versioned replica is provided to
establish each regional production establish each regional production
database. database.
• Regional editors work from their • Regional editors' ArcGIS Desktop
individual production geodatabases. applications are hosted on centralized
• Regional sites reconcile and post their terminal server farm.
edits before synchronizing with • Regional editors work from their
corporate. individual production geodatabases in
• Regional updates are synchronized with the central data center.
the central parent corporate • Regional child production databases
geodatabase. reconcile and post their edits before
• Corporate reconcile and post corporate synchronizing with parent geodatabase.
and regional updates. • Regional updates are synchronized with
• Corporate updates are synchronized the parent geodatabase.
back to the regional geodatabases. • Corporate editors reconcile and post
regional version updates.
• Corporate updates are synchronized
Warning: A system back-up should be
back with regional geodatabases.
completed after each synchronization. It
is important that all child replicas remain
synchronized with their parent versions Best practice: System back-up is
provided after each synchronization

from a common data center storage • Remote Citrix clients accessing central
repository, to maintain consistency Citrix servers for ArcGIS Desktop
between parent and child database access to central maintenance data
volumes. source.
• One-way incremental replication
Distributed geodatabase architecture updates between maintenance and
provides a highly scalable computing publishing database.
infrastructure without placing high demands • ArcGIS Server web applications and
on any single database or server services published to GIS viewers.
component.
Distributed deployment strategies
Enterprise architecture deployment
strategies • Remote ArcGIS Desktop clients
accessing a local data source replicated
Distributed geodatabase replication from a central maintenance
technology integrates a variety of desktop, geodatabase
mobile, and server solutions into an • Local updates are incrementally
adaptive geospatial communications replicated to central maintenance
environment connecting operations across database
the enterprise and throughout the
community. Publishing geodatabase deployed in the
data center or cloud
• Incremental replicated updates from a

central maintenance database
• Highly scalable published web services
to a variety of GIS viewers
ArcGIS Server mapping services can be

deployed directly from the data center, or
geodatabase replication services can be
used to provide incremental updates to
ArcGIS Server web services maintained
within a private or public cloud hosting
Figure 5.14 Enterprise GIS operations often involve a infrastructure. The Cloud computing
mix of geodatabase replication patterns including infrastructure provides a new adaptive
mobile, publication, and distributed operations.
platform environment for managing high
capacity map publishing services.
Enterprise GIS operations often include a
variety of geodatabase replication functions
as shown in Figure 5.14.
Centralized deployment strategies
• Local desktop clients with direct access

to central maintenance geodatabase
and network services.
• Mobile GIS viewers sometimes
connected to central maintenance
geodatabase.

Federated architecture deployment National data center operations
strategies
• Regional databases are integrated at
the national level maintenance
geodatabase.
• National data is replicated to a
publishing geodatabase for public web
services.
• Cloud platform can be used for
publishing national web services.
Best Practice: Multiple database layers

improve security and control level of
data dissemination.
Note: Additional administrative layers can
Figure 5.15 Federated architecture patterns support be included to address business needs.
community and global business requirements,
providing layers of managed and published data
sources at different levels of granularity. Community deployment strategies
Database and Web technology standards

provide opportunities to better manage and
support user access to a rapidly growing
volume of geospatial data resources.
Federated architectures identified in Figure
5.15 provide better data management,
integrating community and national GIS
operations. Geodatabase replication
services and managed extract, transform,
and load (ETL) processes support loosely
coupled distributed geodatabase
environments. Figure 5.16 Cloud-based applications integrate and
synthesize information from many sources, facilitating
communication and collaboration, and are breaking
Multi-layered data architecture patterns are down barriers between institutions, disciplines, and
becoming more common. Solutions include cultures.
the standard desktop, server, and mobile
patterns for each implementation level. The ArcGIS Platform is building community
relationships that change the way people
Regional data center operations work. ArcGIS provides a platform for
collaboration, sharing, and community
• Some local municipalities share regional analysis that helps us better define and
data centers for maintenance and understand the world.
publishing operations.
• Local municipalities publish Technology is changing the way GIS is
maintenance database layers to a serving our communities:
distribution database of record for
regional sharing. • Online GIS platforms deployed in the
cloud enable real-time location analytics
and spatial visualization for everyone.

• Direct spatial query of big data ArcGIS feature services integrate data from
repositories and on-the-fly location multiple enterprise business systems.
analysis are showing relevant, real-time
activity visualized on a spatial map • ArcGIS Desktop software can connect
display on multiple desktop and mobile to supported database (DBMS) content
devices. for view, query, and analysis operations.
• Configurable applications deliver a • ArcGIS Server can connect to supported
solution framework that is easy to build databases and share tabular data as
and maintain for integration, feature services.
collaboration, and efficient information • ArcGIS Desktop supports remote editing
exchange. by creating a local synchronized copy of
• Social media streaming and satellite the data from a published ArcGIS
imagery delivered real-time on filtered Server feature service.
map displays is changing the way • Feature services can be registered with
people understand the world. portal creating feature layer content for
display and analysis in Web maps.
ArcGIS Desktop direct • Web maps can be shared with
authorized users throughout the
connection to supported organization through portal groups and
DBMS content displayed by a variety of mobile and
Web-based COTS applications.
Enterprise geodatabases, also known as

multiuser geodatabases, are stored in a
relational database using DB2, Informix,
SQL Server, Oracle, PostgreSQL. Database
technologies that do not support a
geodatabase can still be leveraged for
accessing spatial related data by the use of
Query Layers. DBMS’s that support Query
Layers include ALTIBASE, Dameng, Azure
SQL Database, Netezza, SAP HANA,
SQLLite, and Teradata.
Figure 5.17 ArcGIS feature services are used to Best practice: ArcGIS Desktop remote
integrate data resources across the organization.
editing workflow can provide an effective
ArcGIS Desktop and ArcGIS Server direct method for editing a centralized
connection to supported database servers database using multiple remote,
supports integration and spatial analysis of sometimes-connected clients.
enterprise data resources. Query layers can
be used to create a database connection,
integrating a live map display with database
tables. Figure 5.17 shows the available
database connection architecture patterns.

Portal distributed collaboration Portal collaboration deployment
patterns
Figure 5.19 Portal distributed collaboration

Figure 5.18 Portal collaboration supports sharing
deployment patterns.
between Portal for ArcGIS sites.
A variety of portal collaboration deployment

Portal collaboration plays a key role in
patterns are supported with the ArcGIS
sharing content between ArcGIS Enterprise
10.5.1 release.
organizations.
Example deployment patterns
Content is shared to or from a group linked
to a collaboration workspace.
• Centralized HUB pattern.
• Collaboration between organizations.
The Host Portal admin defines how Guest
• Integration between an ArcGIS Online
Portals can access each collaboration
and ArcGIS Enterprise organizations.
workspace.
Distributed collaboration options include
• Send - content sent to the collaboration
synchronizing Portal group content and
• Receive - content received from the
copy/synchronizing Portal group feature
collaboration
layer content between ArcGIS Enterprise
• Send and receive – content sent to and
and ArcGIS Online organizations. The
received from the collaboration
option to copy feature layer content
between ArcGIS Enterprise organizations
ArcGIS 10.5.1+ sharing feature layer
will be supported with the ArcGIS 10.6
copies.
release.
(Host feature layer options)
Best Practice: The optimum Portal
• Hosted feature layers in ArcGIS Online
collaboration deployment pattern and
• Hosted feature layers in ArcGIS
synchronization strategy can be
Enterprise
configured to satisfy unique user
• Feature layers registered with portal
collaboration needs.
published from a federated ArcGIS
Enterprise server.

ArcGIS Data Store GIS imagery data
What is ArcGIS Data Store? architecture
Figure 5.20 ArcGIS Data Store functional

configurations.
Figure 5.21 GIS imagery data architecture includes

Base ArcGIS Enterprise configuration ArcGIS Online, pre-processed imagery files, raw
imagery files managed by the mosaic dataset, image
• Portal for ArcGIS service cache, and historical imagery archives
(history).
• Federated GIS Server (configured as
hosting server)
• ArcGIS Data Stores (accessed through Figure 5.21 shows the data source
hosting server) architecture patterns available to manage
imagery data. Imagery data sources include
ArcGIS Data Stores configurations aerial photography, elevation data, and
satellite imagery.
• Relational Data Store
• Tile Cache Data Store Imagery is becoming the most valuable real-
• Spatiotemporal Big Data Store time business intelligence. Aerial
photographs and satellite images can be
collected in real time during a regional
ArcGIS Data Store utilization
emergency and used to evaluate and
respond to national disasters in a timely
• Relational Data Store. Portal feature way. The volume of imagery is growing
services, ArcGIS Image Server, and exponentially as technology for data
ArcGIS GeoAnalytics Server. collection and storage is rapidly evolving to
• Tile Cache Data Store. Portal 3D leverage these digital information products.
Scenes and ArcGIS Image Server.
• Spatiotemporal Big Data Store.
Best practice: Rapid imagery collection
GeoEvent Server and ArcGIS
and publication timelines are key to
GeoAnalytics Server.
proper coordination and response to
natural and man-made national
Base ArcGIS Enterprise configuration is disasters.
required for ArcGIS Data Store deployment.
Imagery can provide valuable information
when information products are examined
over time. Views of the global ice caps and
effects on ground cover can show
information for managing climate change.
Community development and agriculture

changes can benefit from national imagery What is a mosaic dataset?
datasets showing changes in communities
and farm products on a global scale.
An imagery file share is used to store the

raw imagery files.
• Imagery can be stored and delivered

when needed, maintaining optimum
source quality.
• Single preprocessed images can be
distributed using the ArcGIS Server
image service (without the Image
extension). Figure 5.22 A mosaic dataset is a set of tools and a
• Multiple raw imagery files can be metadata catalog for managing a GIS imagery
distributed using a mosaic dataset with repository.
the ArcGIS Server Image extension.
Figure 5.22 shows a mosaic dataset. The
Imagery is a primary resource for visualizing mosaic dataset was developed to manage
live global assets. and deploy imagery information products in
a time-sensitive workflow environment.
• The mosaic dataset provides the core
functionality to organize and manage [A mosaic dataset] consists of:
your imagery data resources.
• ArcGIS Image Server provides Web • A catalog that provides the source of the
client access to imagery data sources pixels and footprints of the rasters.
through a standard REST API. • A feature class that defines the
• The mosaic dataset with the ArcGIS boundary.
Imagery Server enables on-the-fly • A set of mosaicking rules that are used
processing, including multi-image to dynamically mosaic the rasters.
mosaicking as it is compiled through the • A set of properties used to control the
Image Server dynamic service. mosaicking and any image extraction.
• An imagery service cache is a collection • A table for logging during data loading
of preprocessed tiled map images and other operations.
stored at multiple scales for rapid • Optionally, a seam line feature class for
dissemination. seam line mosaicking.
• Imagery historical archiving provides • Optionally, a color correction table that
online access to imagery resources defines the color mapping for each
showing change over time. Hardware raster in the raster catalog.
vendor content-addressable storage
(CAS) solutions provide protected long
term storage for imagery file
repositories.

ArcGIS image access patterns mosaicked as a raster dataset for direct
access by ArcGIS Server Image Service
(without the Image Extension). An imagery
raster dataset (providing single image
access) can be stored in a file system or
loaded into a geodatabase.
Warning: Imagery raster datasets tend to

be quite large (can occupy several
terabytes). Copying large volumes of
raster data into a geodatabase
challenges even the most experienced
database administrators.
ArcGIS Image Server dynamic access to

imagery
Figure 5.23 Imagery can be accessed from ArcGIS
Desktop, ArcGIS Image Server dynamic services The ArcGIS Image Server enables Image
(preprocessed and on-the-fly processing), and ArcGIS
Image Cache. Service access to multiple imagery files.
The ArcGIS Image Server is a licensed
Figure 5.23 shows a variety of available ArcGIS Server role, which extends the
ArcGIS imagery deployment patterns. capability of serving raster data.
Imagery deployment patterns include Specifically, it allows you to use the mosaic
ArcGIS Desktop direct access to imagery, dataset and on-the-fly imagery processing
ArcGIS Server image service access to through the ArcGIS Image Server dynamic
single preprocessed imagery raster service.
datasets, ArcGIS Image Server access to
multiple imagery files with on-the-fly Best practice: With the ArcGIS Image
processing, and direct access to Server, you can serve an imagery
preprocessed imagery cache tiles. repository through a mosaic dataset or
raster dataset layers.
ArcGIS Desktop direct access to imagery
The ArcGIS Image Server gives you the
ArcGIS Desktop provides direct access to ability to:
imagery resources. ArcGIS Desktop is used
to create a mosaic dataset and perform • Put your valuable imagery to use
imagery analysis. Mosaic dataset can be quickly.
used as a common search engine for • Serve collections of imagery or lidar
organizing and accessing local imagery data as image services.
resources. Creating, editing, or working with • Dynamically create and serve mosaics
imagery is a core ArcGIS Desktop capability from the original imagery, without the
and does not require an ArcGIS Image need to pre-compute the mosaics.
extension. • Serve multiple views using the original
imagery.
ArcGIS Server image service access to • Access the catalogs of imagery that
imagery make up the mosaic dataset.
• Exploit overlapping imagery, perform
on-the-fly image processing, and
ArcGIS Server capabilities include Image
explore temporal changes, using the
Service access to pre-processed imagery.
Imagery must be preprocessed and

advanced image-serving capabilities of
this extension.
ArcGIS Imagery cache tiles
ArcGIS Server provides direct access to

preprocessed imagery cache. The image
service cache is a preprocessed pyramid of
imagery tiles configured at a range of
scales. [Image service caching] improves
the performance of image services in client
applications. When accessing cached
imagery using the enable cache view mode,
the preprocessed cached tiles are sent to
the client without future processing.
When you add an imagery cache with an

image service, you end up with a dual-
purpose image service that is accessed
Figure 5.24 The image service cache is preprocessed
depending on its purpose. One purpose is for rapid display performance. Image resources are
to provide the fastest access to the image available with the same image service
as a tiled service. The other purpose is to
provide access through the mosaic dataset Figure 5.24 shows the optimum imagery
to the imagery repository, for queries, service configuration for optimum web
downloading, access to individual items, delivery.
and to use in processing and analysis. Both
options are available through a single image Recommended image caching workflow:
service starting with the ArcGIS 10.1
release. • Create mosaic dataset.
• Serve image services to key users
Specific benefits of a cached image service (dynamic).
include: • Create map cache for larger web
community.
• Improved performance for basemap • Maintain mosaic dataset.
images. • Update cache.
• Skips overview generation.
• Improved performance for slow formats. Image service cache:
Best practice: If your image service is • Provides static background image.

being used as a basemap image (like a • Delivered as tiles for web caching.
map service to serve an image or as a • Most scalable web delivery.
background image), without expecting • Created and served using mosaic
users to modify any of the properties of dataset with ArcGIS Image Server.
the image service, such as changing the • Preprocessing and on-demand caching
mosaic methods, or performing a query, options.
then caching is recommended for
improved performance and scalability. Best practice: When you cache an image
service, you end up with a dual-purpose
Imagery deployment workflow image service that is accessed
depending on its purpose. Caching is

only required when you must create the Figure 5.25 GIS enterprise data architecture is no
fastest possible service containing longer just a geodatabase, it often includes a
combination of both the GIS feature data and the
image data. Generally, the pyramids imagery data resources supporting common services
generated for raster datasets or the managed by an integrated data center operations.
overviews generated for mosaic datasets
result in image data being served at an Figure 5.25 shows the GIS enterprise data
acceptable rate. However, if you know a architecture. GIS enterprise architecture
particular image or area of interest will often includes both feature and imagery
be repeatedly visited, you may want to data within the data center, requiring
generate a cache. effective data management and automation
to maintain the variety of data sources. The
CPT Platform Capacity Calculator GIS data administrator must manage a
custom imagery services hybrid architecture containing a mix of
resources stored on file systems and
The CPT Platform Capacity Calculator can multiple database platforms. ArcGIS
be used to demonstrate variation in Image technology provides a variety of processing
service performance due to the selected and replication functions for maintaining
data source format. data resources in an optimum configuration.
Select a custom imagery workflow Your optimum data configuration will

configuration on the CPT Platform Capacity depend on your business needs.
Calculator tab
Best practice: Data architecture
CPT Platform Capacity Calculator can be solutions are unique to each business
configured to show performance of the operation. Review your user
seven (7) available Imagery data source requirements and operational needs to
formats. establish the optimum data architecture
for your business.
Selecting an imagery workflow on the CPT
Calculator tab
Selecting the imagery workflow on the CPT
Design tab
Storage architecture
GIS enterprise data options
architecture
Storage technology has evolved over the
past 20 years to improve data access and
provide better management of available
storage resources. Understanding the
advantages of each technical solution will
help you select the storage architecture that
best supports your needs.
Advent of the storage area network

and others with too little. A more adaptive
storage management solution was needed.
Direct attached storage (DAS)
DAS moves the primary storage volumes

into a separate platform tier with data
volumes that can be assigned to servers as
required to satisfy data storage needs.
Figure 5.26 The storage area network evolved to • Server host bus adapters (HBA) and
satisfy adaptive storage management needs for data fiber channel communications
centers with many database servers.
maintained the same Small Computer
System Interface (SCSI) communication
Figure 5.26 shows the evolution of the protocol used for internal storage.
storage area network. A storage area • Multiple fiber channel connections were
network provides an optimum data provided to connect a single storage
management solution for a data center with array with multiple database servers.
many database servers. • Disk volumes were allocated as required
to each database server and were not
Local disk storage shared.
• Storage arrays were configured with
Local disk storage is provided for desktop redundant components to support
workstations, laptops, and mobile devices. operational high availability
requirements.
Best practice: Optimum workstation
configurations today include two local As the data centers grew, there was a
disks for enhanced display performance demand for more ways to allocate storage
and data protection. volumes to the growing number of database
server platforms.
Internal disk storage
Storage area networks (SAN)
Internal disk storage is provided in file
servers following the same pattern used for SAN establish a network between the
desktop workstations. multiple database servers and the multiple
storage arrays providing adaptive
• Many enterprise business solutions, connectivity for assigning storage volumes
including GIS, moved their data to the server platform tier as required to
resources to database management meet operational needs.
systems in the 1990s.
• Database servers were initially • The initial SAN switches provided fiber
purchased with larger internal storage channel port connections compatible
bays to accommodate higher capacity with existing HBA and storage array
data storage volumes. cabling.
• SAN establishes a fiber channel network
Internal disk storage architecture started to for routing the storage traffic.
cause a problem for larger data centers, as • Any storage volume on the storage
they found their disk storage assets would array tier could be assigned to any
be silos of dedicated storage in server database platform on the server tier.
platforms, some with too much disk capacity

Best practice: SANs have became an As NAS technology started to become
optimum storage solution for large data popular (late 1990s), most enterprise
center environments with many database business solutions, including GIS, were
servers. moving their data to database management
systems. The SAN solution was winning
Advent of network-attached storage over the NAS for most large data center
deployments.
Figure 5.27 shows the evolution of network
attached storage. Network attached storage Simple Computer Storage Interface over
provides an optimum data management IP (iSCSI)
solution for high-capacity file-based storage
(high-availability file share appliances). iSCSI was developed for building SAN
Direct-attached storage provided the first solutions from NAS appliance systems
business case in favor of network attached (iSCSI storage area network).
storage appliances.
• Supports transmission of the traditional
SCSI block protocol over IP networks.
• Provides database vendors dedicated
storage volumes
• IP switch bandwidth also increased to
higher capacity than what was achieved
by the fiber channel switches, helping
with the performance concerns (10
Gbps IP switch technology).
• The NAS appliances architecture can
provide an adaptive solution for both IP
Figure 5.27 Network attached storage was file sharing and iSCSI database storage
established as an optimum solution for high- management.
availability file share environments.
Fiber Channel over Ethernet (FCoE)
Direct-attached storage (DAS) had disk
storage volumes dedicated to a specific FCoE was developed to provide SCSI block
assigned server. Two clustered file servers protocol over standard Ethernet networks
were required to create a high-availability without the IP routing overhead.
file share. This DAS file share pattern was
expensive and difficult to manage. • Like iSCSI, FCoE uses standard
multipurpose Ethernet networks to
Network-attached storage (NAS) connect servers with storage.
• Unlike iSCSI, FCoE does not run over
NAS provided a high-availability file share TCP/IP (less processing overhead) - it
appliance which would simply connect to a uses its own Ethernet protocol.
standard Internet Protocol (IP) local area • Advantage, FCoE runs over the same
network switch (file share protocols and full general purpose switches that iSCSI
high availability redundancy were included does.
with the NAS appliance). The NAS • Limitations, FCoE cannot be routed over
appliance provides a simple plug-and-play a TCP/IP WAN (limited to stable local
solution for including a file share on a local network communications).
area network. • Requires Fiber Channel administration
and management (more complex that
iSCSI).

Best practice: NAS provides an optimum disk can be replaced and data restored
storage architecture for enterprise GIS automatically from the mirror copy without
operations. Most enterprise GIS data bringing the storage array down for
centers include a mix of SAN (FC, FCoE, maintenance. RAID 1 provides optimum
or iSCSI) and NAS solutions to satisfy data protection with minimum performance
their data management needs. gain. Available data storage is limited to 50
percent of the total disk volume, since a
RAID (Redundant array of mirror disk copy is maintained for every data
independent disks) disk in the array.
Enterprise GIS environments depend RAID 3 and 4: A disk volume in a RAID 3 or

heavily on GIS data to support a variety of RAID 4 configuration supports striping of
critical business processes. Data is one of data across all disks in the array except for
the most valuable resources of a GIS, and one parity disk. A parity bit is calculated for
protecting data is fundamental to supporting each data stripe and stored on the parity
critical business operations. disk. If one of the disks fails, the parity bit
can be used to recalculate and restore the
The primary data protection line of defense missing data. RAID 3 provides good
is provided by the storage solutions. Most protection of the data and allows optimum
storage vendors have standardized on use of the storage volume. All but one parity
redundant array of independent disks disk can be used for data storage,
(RAID) storage solutions for data protection. optimizing use of the available disk volume
A brief overview of basic storage protection for data storage capacity.
alternatives includes the following:
There are some technical differences
Just a Bunch of Disks (JBOD): A disk between RAID 3 and RAID 4, which, for our
volume with no RAID protection is referred purposes, are beyond the scope of this
to as just a bunch of disks configuration, or discussion. Both of these storage
(JBOD). This represents a configuration of configurations have potential performance
disks with no protection and no performance disadvantages. The common parity disk
optimization. must be accessed for each write, which can
result in disk contention under heavy peak
user loads. Performance may also suffer
RAID 0: A disk volume in a RAID 0
because of requirements to calculate and
configuration provides striping of data
store the parity bit for each write. Write
across several disks in the storage array.
performance issues are normally resolved
Striping supports parallel disk controller
through array cache algorithms on most
access to data across several disks
high-performance disk storage solutions.
reducing the time required to locate and
transfer the requested data. Data is
transferred to array cache once it is found The following RAID configurations are the
on each disk. RAID 0 striping provides most commonly used to support database
optimum data access performance with no storage solutions. These solutions represent
data protection. One hundred percent of the RAID combinations that best support data
disk volume is available for data storage. protection and performance goals.
RAID 1: A disk volume in a RAID 1

configuration provides mirror copies of the
data on disk pairs within the array. If one
disk in a pair fails, data can be accessed
from the remaining disk copy. The failed

Best practice: GIS feature data and
imagery files can be located on striped
RAID 5 disks with minimum performance
impact.
Figure 5.28 RAID 1/0 provides optimum protection

and optimizes query performance.
Figure 5.28 shows the RAID 1/0 storage

disk layout. RAID 1/0 is a composite
solution including RAID 0 striping and RAID Figure 5.30 RAID 6 provides better protection and
better query performance.
1 mirroring.
Figure 5.30 shows the RAID 6 storage disk
• Optimum solution for high performance
layout. RAID 6 supports striping of data
and data protection.
across all disks in the array except for two
• Highest cost solution. Available data
parity disks.
storage is limited to 50 percent of the
total disk volume, since a mirror disk
copy is maintained for every data disk in • If one or two disks fail, the parity bit can
the array. be used to recalculate and restore the
missing data.
• RAID 6 provides improved protection
Best practice: High-activity database
and reduced data contention (larger disk
index tables and log files are best
volumes at same protection levels) than
located on RAID 1/0 storage volumes.
RAID 5.
• Primary driver for RAID 6 was the longer
rebuild times required for the larger
volume disks (higher risk for two-disk
failure).
Best practice: Use RAID 6 to reduce

concern of data loss due to a two-
Figure 5.29 RAID 5 provides optimum protection and concurrent disk failure scenario.
good performance.
Will storage be the next performance
Figure 5.29 shows the RAID 5 storage disk bottleneck?
layout. RAID 5 supports striping of data
across all disks in the array except for one Technology is improving display
parity disk. performance and moving more data faster
and more efficiently throughout the server,
• A parity bit is calculated for each data network, and storage infrastructure.
stripe and stored on the parity disk.
• If one disk fails, the parity bit can be • Larger disk volumes (reduced number of
used to recalculate and restore the disks in array)
missing data. • Increasing storage traffic loads (cached
• Provides optimum disk utilization and tiles, imagery files)
near optimum performance, supporting • Faster display processing
disk storage on all but one parity disk • Higher capacity servers
volume. • Larger peak concurrent GIS user loads

• Virtual Server and Virtual Desktops HDD impacts on display performance vary
deployed on high capacity host based on workflow data access patterns,
platforms the location of data on the disk, and the disk
rotation speed. HDD disk rotation speed
Warning: All technology advances point impacts the capacity of the drives and the
toward increased potential for disk cost. Higher capacity drives are available at
contention. the lower rotation speeds delivering a
significantly lower cost per GB storage. The
The good news is that there are technical number of disk in a RAID storage volume
solutions available to resolve contention. It also impacts performance (more disk
is also quite simple to monitor disk I/O enable higher parallel throughput
performance and identify if disk contention performance).
is a problem. Available solutions to disk
contention include: A solid state drive (SSD), on the other hand,
has no moving parts and is capable of
• Cache data files on edge servers or web accessing any location on the drive with
accelerator appliances (servers located equally fast speed and precision. SSD read
near the site communication access and write performance can be 100 times
points). faster than HDD, with a much smaller form
• Disk contention can also be reduced factor and much less power consumption.
with RAID data stripping – distributing Samsung Solid State Drive White Paper
the data across larger RAID volumes provides a very helpful overview of the
can reduce the probability of disk current state of SSD technology and what
contention. you need to know when selecting the right
• Solid-state disk drives are available in storage technology solution for your
the current marketplace; solutions that environment.
deliver data over 100 times faster than
current mechanical disk drives. Flash memory technology
Moving to solid state storage Solid state storage technology started out
technology as flash memory. All Flash devices have
certain basic properties in common:
• Non-volatile storage of code and data is

stored in units called “pages”
• Groups of pages belong to “erase units”
known as “blocks”. A block is the
smallest SSD sector that can be erased.
• Storage blocks must be erased before
pages can be written to.
• Data storage blocks will wear out after a
Figure 5.31 Solid State Technology is replacing certain number of erase/write cycles.
mechanical drives as a foundation for high
performance operations and mobile deployment.
There are two basic types of flash chips:
A Hard Disk Drive (HDD) works by way of a NOR and NAND (named after the NOR and
mechanical drive head that must physically NAND logical programming gates). NOR
move to access locations on a rapidly- Flash chips support random access with
spinning magnetic disk. Rotating disks must execute-in-place capability and is commonly
wait for spindle motors, heads, and arms to used to run code (direct read only). NAND
physically locate data sectors. Flash chips can store approximately four

times as much data as NOR for the same Enhanced SSD controller capabilities have
price, delivers much faster erase and write advanced MLC cell endurance through a
times, and is the chip of choice for Solid variety of amelioration techniques creating
State Drive technology. an attractive enterprise (eMLC) chip
configuration. Amelioration techniques
Solid State Drive technology include:
• Wear-leveling: moves write cycles

around the chip so that cells wear
evenly.
• On-device deduplication: reduces the
volumes of data written and so lowers
wear.
• Redundancy: reserves a portion of the
device’s capacity to replace cells as
they fail.
• Write optimization: stores data writes so
they can be made in large chucks to
Figure 5.32 Solid State Drive technology advances reduce the number of erase/write
deliver enterprise class storage in a competitive operations.
marketplace.
There are three classes of NAND SSD The eMLC SSD configuration significantly
chips. SSD classes are defined by the reduces the cost of storage by providing
number of electrical charges that are stored SLC class endurance (100,000 erase/write
in each NAND cell. cycles) at the MLC class capacity (4 times
SLC class). Hitachi shares their release of
• Single-Level Cell (SLC)]: 1 bit of data SLC NAND Flash and eMLC NAND Flash
per cell Enterprise-class SSDs.
• Multi-Level Cell (MLC)]: 2 bits of data
per cell Hierarchical storage implementation
• Three-Level Cell (TLC): 3 bits of data
per cell
The more bits a cell stores at one time, the

more capacity that fits in one place reducing
manufacturing costs and increasing SSD
capacity. Manek Dubash shares his review
on MLC vs SLC: Which flash SSD is right
for you?
SSD endurance (maximum erase cycles) Figure 5.33 Hierarchical storage architecture
introduces solid state storage as an integrated
varies based on technology class. Multi- extension of existing storage solutions.
Level Cell configurations are increasingly
sensitive to electric charge deterioration Most of the enterprise storage market today
(wear out). is supported by HDD technology. SSD
technology will not replace spinning disks
• SLC: 100,000 erase/write cycles any time soon – so many of the popular
• MLC: 10,000 erase/write cycles enterprise storage solutions involve
• TLC: 1,000 erase/write cycles implementation of a hierarchical storage
architecture. With hierarchical storage, only

the working data set needs to be on SSD -
and typically that’s about 5 to 15 percent of
the total on-line data repository.
Most new laptops and mobile devices are

now supported by SSD technology. Intel is
starting to include SSD storage on PCI ports
with the computer motherboard, enabling Figure 5.34 Disk or tape backup remain the most
optimum bandwidth access to critical popular ways to move large volumes of GIS data.
cached data sources. A variety of storage
vendors are providing SSD based gateway Figure 5.34 shows ways to move large
products that connect to existing SAN and datasets. Large volumes of data are best
NAS storage, storing a cached copy of moved on disk, DVD, or tape back-up.
working business data on eMLC SSD
components with historical resources Moving large volumes of data across shared
retained on existing HDD storage. These network resources is very expensive, and
hierarchical storage solutions deliver SSD impacts performance for all users on that
class operational performance gains while network segment. Network bandwidth is a
continuing to leverage existing investment in valuable and limited resource. If you
HDD technology. conserve network bandwidth for primary
display and query tasks, you will have more
Best practice: capacity available for better display
1) Be aware of the potential for performance.
performance bottlenecks within your
storage environment.
2) Monitor disk I/O performance to
identify when disk contention is causing
a performance delay. Protect your GIS data
3) Be aware that there are solutions to
disk I/O performance problems, and take resources
appropriate action to address
performance issues when they occur. Backup data center
4) Consider Solid State Disk storage
solutions for future technology
investments.
Ways to move GIS data

GIS data can be moved using an extract
and load process or by forms of replication.
The optimum method for moving your data
will be determined by the amount of data to
Figure 5.35 Continuity of Operation Planning (COOP).
be moved and the associated business
processes.
Many organizations have a requirement to
maintain continued operations under all
Traditional tape backup/disk copy
possible disaster recovery situations.
Normally this requires redundant multiple
data center operations.

Multiple synchronized data center Warning: ArcSDE geodatabase
operations require special architecture replication should not be used for back-
considerations. up and recovery operations.
• Multiple-site install and configuration Disk-level replication

• Multiple-site service deployments
• Multiple-site data source updates
• Concurrent application deployment and
updates at both locations
Consider the cost and benefits when

requiring a COOP solution.
• Do you need two data centers with Figure 5.37 Storage-level replication provides back-up
to a separate storage location.
immediate failover?
• How much data loss would be
reasonable and acceptable for Figure 5.37 shows a storage replication
continued operations? architecture. Storage disk-level replication
• Can an off-site backup provide provides the optimum solution for data
adequate business protection? center back-up and recovery operations.
Best practice: Planning for failure before Storage vendors typically provide
the disaster occurs can make a life and incremental snapshot back-ups to local and
death difference in maintaining remote data center locations. Many of these
continuity of operations. solutions include proven recovery tools and
are widely in use. Back-up from storage
volumes avoids database complexity
Database replication
issues, data is replicated at the disk storage
block-level.
Best practice: Use storage disk

replication when you want to create and
maintain a complete copy of the data
storage environment. Storage vendors
provide optimum tools for data
protection, back-up, and recovery
Figure 5.36 Database replication is a good solution
when you are keeping a replicated copy of all the
operations.
database transactions.
Figure 5.36 shows a database replication

architecture. Commercial database
replication works well for maintaining a real-
time back-up failover database.
Best practice: Use database replication

when you want to create and maintain a
complete copy of the database
environment. Database vendors provide
optimum tools for data protection, back-
up, and recovery operations.

Data Backups some of the most probable situations being
administrative or user error.
Host Tape Backup: Traditional server

backup solutions use lower-cost tape
storage for backup. Data must be converted
to a tape storage format and stored in a
linear tape medium. Backups can be a long
drawn out process taking considerable
server processing resource (typically
consume a CPU during the backup process)
and requiring special data management for
Figure 5.38 Snapshot backup can protect your data
investment.
operational environments.
Data protection at the disk level minimizes For database environments, point-in-time
the need for system recovery in the event of backups are required to maintain database
a single disk failure but will not protect continuity. Database software provide for
against a variety of other data failure online backup requirements by enabling a
scenarios. It is always important to keep a procedural snapshot of the database. A
current backup copy of critical data copy of the protected snapshot data is
resources, and maintain a recent copy at a retained in a snapshot table when changes
safe location away from the primary site. are made to the database, supporting point-
Figure 5.38 highlights data backup in-time backup of the database and
strategies available to protect your business potential database recovery back to the time
operations. It is important to maintain a of the snapshot.
snapshot back-up or copy of your data. A
large percentage of data loss is caused by Host processors can be used to support
human error. The best way to protect your backup operations during off-peak hours. If
data is to maintain a reliable periodic point- backups are required during peak-use
in-time back-up strategy. periods, backups can impact server
performance.
The type of backup system you choose for
your business will depend on your business Network Client Tape Backup: The
needs. For simple low priority, single use traditional online backup can often be
environments, you can create a periodic supported over the LAN with the primary
point-in-time backup on a local disk or tape batch backup process running on a
drive and maintain a recent off-site copy of separate client platform. DBMS snapshots
your data for business recovery. For larger may still be used to support point-in-time
enterprise operations, system availability backups for online database environments.
requirements may drive requirements for Client backup processes can contribute to
failover to backup Data Centers when the potential network performance bottlenecks
primary site fails. Your business needs will between the server and the client machine
drive the level of protection you need. because of the high data transfer rates
during the backup process.
Data backups provide the last line of
defense for protecting our data investments. Storage Area Network Client Tape
Careful planning and attention to storage Backup: Some backup solutions support
backup procedures are important factors to direct disk storage access without impacting
a successful backup strategy. Data loss can the host DBMS server environment. Storage
result from many types of situations, with backup is performed over the SAN or

through a separate storage network access The question is not if, but when. Most
to the disk array with batch process running people will, at some time, experience loss of
on a separate client platform. A disk-level valuable data and business resources.
storage array snapshot is used to support
point-in-time backups for online database
environments. Host platform processing
Data Management Overview
loads and LAN performance bottlenecks
can be avoided with disk-level backup Support for distributed database solutions
solutions. has traditionally introduced high-risk
operations, with potential for data corruption
and use of stale data sources in GIS
Disk Copy Backup: The size of databases
operations. There are organizations that
has increased dramatically in recent years,
support successful distributed solutions.
growing from tens of gigabytes to hundreds
Their success is based on careful planning
of gigabytes and, in many cases, terabytes
and detailed attention to their administrative
of data. Recovery of large databases from
processes that support the distributed data
tape backups is very slow, taking days to
sites. More successful GIS implementations
recover large spatial database
support central consolidated database
environments. At the same time, the cost of
environments with effective remote user
disk storage has decreased dramatically
performance and support. Future distributed
providing disk copy solutions for large
database management solutions may
database environments competitive in price
significantly reduce the risk of supporting
to tape storage solutions. A copy of the
distributed environments. Whether
database on local disk, or a copy of these
centralized or distributed, the success of
disks to a remote recovery site, can support
enterprise GIS solutions will depend heavily
immediate restart of the DBMS following a
on the administrative team that keeps the
storage failure by simply restarting the
system operational and provides an
DBMS with the backup disk copy.
architecture solution that supports user
access needs.
There are several names for disk backup
strategies (remote backup, disaster
recovery plan, business continuance plan, CPT Capacity Planning videos
continuation of business operations plan,
etc). The important thing is that you The next chapter will discuss Network
consider your business needs, evaluate Communications, providing some insight on
risks associated with loss of data resources, how to build GIS solutions that support
and establish a formal plan for business remote user performance and Enterprise
recovery in the event of data loss. GIS scalability.
Page Footer
System Design Strategies 26th edition - An Esri ® Technical Reference Document • 2009 (final
PDF release)

Network Communications
Fall 2017 Network Communications 41st Edition
Network communications provide the conferencing. GIS technology provides an

required connectivity for distributed GIS intuitive visual display to the user, enabling
operations. Network capacity, in many review of lots of maps for high user
cases, can limit the software technology productivity. Each GIS map includes lots of
solutions that perform well within your data accessed from a variety of shared data
organization. System architecture design sources enabling real-time display and
must identify and address network analysis. Data must be transported to where
communication constraints and provide the the program is executed to display the
right technical solution for a successful GIS information, generating lots of traffic across
implementation. the network.
Communication technology advances are What GIS does

promoting a rapid shift in GIS client
technology patterns, providing worldwide • Displays data in a user-friendly graphic
access to millions of data sources from a representation (maps).
variety of desktop and mobile client devices • GIS map display represents overlay of
in real-time. many layers of data resources.
• User quickly sees what is important and
These technology solutions work well if you requests another map display (up to 10
have the required network connectivity. displays per minute).
These solutions will not work if you do not • GIS data is typically shared from a
have the required network connectivity. This central data repository.
chapter will show how to build a GIS that • Data layers are accessed from the
will perform well within the constraints of central data repository to complete each
your network communications infrastructure. client map display.
• User productivity (viewing lots of maps)
generates a lot of network traffic.
Why is GIS traffic- System architecture design should include
intensive? an evaluation of bandwidth capacity
(network suitability analysis) to ensure
that peak traffic loads can be
accommodated to satisfy user performance
requirements.
Warning: Network bandwidth limitations

(not enough bandwidth to handle peak
traffic flow) can seriously constrict user
performance during peak operational
loads.
Figure 6.1 GIS users’ workflows generate relatively
high levels of traffic across the network.
GIS applications rate among the heavy

users of network traffic, along with
document management and video
Network Communications 6.1

Types of networks inexpensive relative to other hardware costs
supporting the system environment.
Network communication technology can be
• Relatively short communication
grouped into two general classes. These
distances (within buildings or
technology classes include LANs and
campus environments)
WANs. the capacity of a particular network
• Relatively high bandwidth capacity
segment is identified by the maximum data
• Relatively low-cost infrastructure,
transfer (measured in bits) that can be
compared to alternative wide area
transported per second passing through that
networks
device. This capacity is called network
bandwidth and is typically measured in
millions of bits [megabits (Mb)] or billions of Best practice: Local area network design
bits [gigabits (Gb)] per second. requirements can normally be addressed
through recommended bandwidth
specifications generated from the
Local area networks (LAN) system architecture design network flow
analysis.
Wide area networks (WAN)
Figure 6.2 Local area network connects client and

server computers over a building or campus
environment. Organizations normally purchase and
maintain the network hardware and cabling
Figure 6.3 Wide area network connects remote client
infrastructure for their own local area networks)
and server computers to a data center located at a
different geographic location. Organizations normally
Figure 6.2 shows a drawing representing a rent bandwidth from a commercial network service
local area network. Local area networks provided who provides the communication
provide high-bandwidth communications for infrastructure for WAN connections)
short distances. These networks connect
local operating environments (usually within Figure 6.3 shows a drawing representing a
a building or local campus environment). wide area network. Wide area networks
Data transport over a single technology provide an infrastructure enabling
medium is single-threaded, which means communication between remote locations.
only one data transmission can be WAN connections normally provide much
supported on a single LAN segment at any lower bandwidth than LAN environments,
time. The cost for LAN environments is but transmission is possible over long

distances. The cost for WAN connections is The bandwidth you have will determine the
relatively high compared to LAN amount of data you can transport.
environments. The Internet is a global WAN
environment. • Data is a collection of digital computer
information stored in media that have
• Relatively long communication the capability to record and retain the
distances between remote locations data structure.
• Relatively low bandwidth capacity • Data can be transported from one
• Relatively high-cost infrastructure, location to another over different types
compared to alternative local area of media.
network analysis • Each network protocol has limits to its
supported rate of data transport, based
Best practice: Careful evaluation of on the technology used to support the
remote site network bandwidth capacity transmission.
is critical during a system architecture • Network traffic capacity is measured in
design analysis. Limited infrastructure terms of connection bandwidth.
capacity can be a primary factor in o GIS uses a lot of data, and GIS
selecting the appropriate software users normally identify data in
technology solution. Megabyte units.
o Network bandwidth is based on
Warning: Many GIS implementations
Megabit units.
have failed due to inadequate
accommodation of peak WAN traffic
flows. Remote client ability to use their 1 Megabyte (MB) = 8 Megabits (Mb).
GIS applications will depend heavily on
having adequate network capacity to Data capacity is measured in terms of
communicate with required GIS data megabytes or gigabytes when stored on
resources. computer disk. Megabyte is abbreviated
using a large "B," while megabit is
abbreviated using a small "b." One must
What is network capacity? remember 1 MB = 8 Mb when converting
data volume from disk storage to network
GIS software solutions continue to generate traffic. Network traffic also includes some
more and more traffic across the network protocol overhead, so a simple rule-of-
infrastructure with each release - and this is thumb is to translate 1 MB of data to about
not the only technology making increasing 10 Mb of traffic.
demands on available network bandwidth.
Bandwidth is often a limited resource. It is Warning: Bandwidth capacity units are
important that we conserve shared network important. Make sure you have the right
resources, and select technology units.
deployment patterns that make productive
use of this limited resource. We can make a
difference by the way we work and how we
leverage available network resources in our
technology selections.
Network capacity is identified as

bandwidth.

LAN bandwidth capacity Warning: Wireless connections are less
stable than wired connections and often
achieve much lower capacity than shown
in their published specifications.
WAN bandwidth capacity
Figure 6.4 Local area network providers have

standardized on Ethernet protocols for wired local
networks. Wireless communications is becoming
more popular with the rapid growth of mobile
technology
Figure 6.4 shows the local area network

protocols and available bandwidth capacity
with current technology offerings. Figure 6.5 Wide area network service
providers use a wide variety of protocols for
Ethernet connections: worldwide connectivity.
• Ethernet connections are provided over Figure 6.5 shows the wide area network
twisted pair cable or optical fiber direct protocols and available bandwidth capacity
wired connections. with current technology offerings.
• Connection bandwidth capacity is
limited by network switch or router Note: Wide area networks use a broad
device specifications. variety of protocols.
• Cost of the network switch or router
device is based on the provided Telephone lines and Internet infrastructure:
connection capacity.
• Dial-up modem
Best practice: Direct wired Ethernet • Telephone service providers offer T-1,
networks provide the most stable T-2, T-3, T4, and T5 communication
network connections. gateways. Slightly higher capacity E1,
E2, E3, E4, and E5 communication
Local wireless connections: gateways are used by European service
providers.
• Local wireless connections are provided • Internet cable infrastructure connections
over a specified range of radio include higher capacity OC3, OC12,
frequencies. OC48, OC192 lines that make up the
• Wireless bandwidth capacity depends service provider cloud infrastructure.
on available technology. • Free space optical and microwave
communications provide alternative

transmissions when optical cable mechanical device that could be trained to
connections are not feasible. do our brain work?
• Satellite communications use
microwave technology.
• laser transmitters send optical signals to
line-of-site receivers over long distance
hops.
• Cable TV infrastructure can be used for
data transmissions, and is able to
provide high bandwidth connections
throughout the provided infrastructure.
Warning: The cost of wide area

bandwidth capacity can be very
expensive and capacity may not be Figure 6.6 Data is a pattern of on and off switches
that represent our thoughts, ideas, and imagination.
available. Make sure to evaluate
bandwidth requirements and availability
Figure 6.6 gives some clues on what data
during your system architecture design
really is. Our thoughts, our ideas, our
assessment.
imagination. Data is a pattern that
represents what we as humans do.
Mobile cell phone network:
How has man communicated in the past?
• Mobile cell phones communicate over
radio frequencies bands, connecting to
• Pictures
the nearest tower transmitters located in
• Drawings
a cellular infrastructure throughout the
• Maps and charts
service area.
• Language (a pattern of symbols that
• Cell phone wireless capacity depends
share our thoughts and ideas)
on the available technology.
So what is data?
Warning: Make sure your smart phone
device and service provider towers in
your operating location satisfy your • Computers represent data in patterns of
mobile bandwidth performance 1s and 0s.
requirements. • These patterns of 1s and 0s are called
data.
• These patterns represent thoughts and
What is data? ideas.
• These patterns can be manipulated to
I remember a trip I made to the Chicago produce information products that can
Science and Industry Museum when I was a be used in your work.
young boy. They had a simple light bulb
display that demonstrated how computers Data is not a physical thing that has mass
think. This was in the early 1960s, when and takes up space—the content is a
computers were not common household pattern representing our ideas. Only the
appliances. The display showed how a medium and method in which data is stored
series of on and off light bulbs could be takes up space. Understanding this concept
used to add and subtract numbers - the is important, since it explains why there are
scientific breakthrough that gave us so many different ways to store and move
computers. Who ever thought we could data.
share our thoughts and ideas with a

Our understanding of data suggests the transmission format and address
physical laws of space and travel do not information.
apply to data, only to the medium that is • Data is packaged for transport across
used to store and move data. This opens a the network.
world of possibilities to communicate and • Packets contain communication control
store data in ways that may not even be information required to transport the
imagined today—humans are limited only data from its source client process to the
by imagination. destination server process.
• Data package is translated to another
There are satellites collecting high- medium for transmission across the
resolution images of the world that can be network.
displayed on a cell phone—an idea beyond • Data package is returned to its original
imagination less than 10 years ago. computer language for final processing.
• Data is unpackaged by the server and
Will science soon overcome the returned to its original state.
communication infrastructure limitations
experienced today? It does seem likely. What are network transport protocols?
Warning: Physical medium limitations

for storing and moving data today can
limit viable GIS technology choices.
What is client/server communication?
Figure 6.8 Network communications involve several

layers of protocol transforms to move data from one
host to another.
Figure 6.8 shows a simple overview of a

network protocol stack. The communication
packet is constructed at different layers
during the transmission process. These
include the application layer, transport layer,
Figure 6.7 Client/server protocols enable Internet layer, and network access layer.
communication of data across the network. Each layer includes sublayers that complete
the data packaging for transport.
Figure 6.7 shows a simple diagram of how
network communications work. Data is A data stream from the host A application is
passed over the network using a variety of sent through the protocol layers to establish
client/server communication protocols. Data a data frame for network transmission.
is packaged for transmission across the
network by the sender machine. Information • The transmission control protocol (TCP)
is provided to aid in routing the data to its header packages the data at the
destination. Data is unpacked at its transport layer.
destination and returned to its original state. • The Internet protocol (IP) header is
added at the Internet layer.
• Communication processes located on
client and server platforms define the

• The media access control (MAC) Best practice: GIS users generally
address information is included at the measure data size in terms of megabytes
physical network layer. or gigabytes. Data size on disk in
• The data frame is then transmitted megabytes (MB) can be converted to
across the network to host B, where the traffic in megabits (Mb) by multiplying by
reverse process moves the data to the a factor of 10 (8 bits of data plus two bits
host application. of overhead for each byte of traffic).
A single data transfer can include several

communications back and forth between the
host applications.
GIS Client/Server Communication Protocols
Figure 6.9 GIS protocols that support communication between the software application and the supported data
source. GIS remote client access communication protocols include Windows Terminal Server (WTS) and web (HTTP)
clients.
Figure 6.9 provides an overview of the Network file services (NFS) and
primary communication protocols used for Common Internet File System (CIFS)
GIS operations. Each protocol Protocols
implementation includes client and server
components to support data delivery. The Support remote disk mounting enabling a
client process prepares the data for client application to access data from a
transmission, and the server process distributed server platform (network file
delivers the data to the application share). All query intelligence resides in the
environment. Roughly 2 MB of data is client application, providing query
needed from the server to complete a instructions to access data located on the
medium-complexity GIS map display. server platform. Data must be transferred to

the client application to support query • Data is uncompressed by the ArcGIS
execution. client application.
• Data must be transferred to the client
Benchmark testing with a remote shape file application to support analysis and
database identified 2 MB per display data display.
transfer (medium-complexity display),
showing a traffic overhead of four times the ArcSDE API (Application server connect
required 2 MB data transfer load. architecture)
• All query intelligence resides in the • Initially SDE was developed to extend
client application, providing query DBMS support for spatial data types.
instructions to access data located on • SDE would be installed on the server for
the server platform. translating ArcGIS query calls for the
• Data must be transferred to the client supported DBMS.
application to support query execution. • A proprietary ArcSDE API was used by
• Chatty communications with shapefile ArcGIS clients to connect to SDE
data source resulting in high traffic installed on the DBMS server.
overhead.
An alternative direct connect option is
Best practice: Most large GIS files available that connects with a DBMS client
include a spatial index. The ArcGIS client application program interface (API)
will first download the spatial index, executed on the client desktop. The ArcSDE
identify the location of the portion of the middleware functions are supported on the
file required from disk, and then request client platform, and the DBMS network
delivery of the portion of the file required client supports data transmission to the
to complete the display (complete file server. Query processing remains on the
deliver would not be required). DBMS server.
Database Access Protocols Database (DB) API (ArcGIS direct connect

architecture)
ArcSDE includes client and server
communication components. The database • ArcGIS client direct connect option
management system (DBMS) includes connects with a DBMS client application
intelligence to support query processing. program interface (API) executed on the
client desktop.
Enterprise geodatabase traffic (medium- • The SDE translation functions are
complexity display) is roughly 2 Mbpd, with supported on the client platform
roughly 50 percent data compression when connecting the DBMS client, and the
stored in an enterprise geodatabase. Data DBMS network client supports data
is decompressed when processed by an transmission to the server.
ArcGIS client. • Query processing remains on the DBMS
server.
• ArcSDE includes client and server • ArcGIS direct connect supports feature
communication components. access to a variety of supported
• The database management system database environments.
(DBMS) includes intelligence to support
query processing. Best practice: Direct connect is the
• Compressed data is provided for recommended database architecture
network transfer and database storage solution.
(roughly 50 percent compression).

ICA and RDP Protocols HTTP Protocols
The Microsoft Remote Desktop Protocol The Hypertext Transfer Protocol (HTTP) is a
(RDP) and Citrix Independent Computing standard Web transmission protocol. In this
Architecture (ICA) protocols support remote transaction-based environment, service
terminal display and control of applications selection and display are controlled by the
hosted on a shared Windows Terminal browser client. Data is compressed for
Server (Remote Desktop Services). transmission. There are a variety of Internet
Protocols that can be implemented within
• The ArcGIS application running in the the HTTP framework.
data center accesses the GIS data and
generates the map display. Web service map images can be embedded
• The display output is delivered to the in a web browser.
user terminal client display.
• The amount of data in the map display • Size of the map image is determined by
output is roughly 200 KB, a factor of 10 server administrator during web
less than the original source data used publishing.
to generate the display. • 100–200KB map image files are
common with each display transaction
Microsoft (RDP) and Citrix ICA protocols: (1 Mb to 2Mb of traffic per display
transaction).
• These protocols support remote terminal
display and control of applications Web service map images support rich
hosted on a shared Windows Terminal Internet application clients.
Server.
• Both protocols transmit display and • Size of the map image is determined by
control information to the terminal client. the resolution of client display.
• Both the Independent Computing • 200–400KB map image files are
Architecture (ICA) protocol and Remote common with each display transaction
Desktop Protocol (RDP) compress data (2Mb to 4Mb of traffic per display
for transmission. transaction)
• Maps from vector-only data layers can
generate less than 28 KB of traffic per Warning: Traffic for ArcGIS for Desktop
display. Compression of common pixel clients is often higher because of the
color areas is very good, resulting in the larger image transfers. Image size is
lower traffic volume. directly proportional to the physical
• Maps that include an imagery layer screen display resolution and map
generate up to 200 KB of traffic per display size; thus, larger image displays
display. Imagery results in color result in higher traffic.
variation between most pixels, limiting
the amount of traffic compression.
Best practice: Most GIS customers

choose to use the Citrix ICA protocol for
hosting centralized ArcGIS for Desktop
operations. Citrix ICA protocol provides
the best user experience and advanced
terminal services administration.

Network Performance • Network transport time = minimum
network gateway processing time
(display traffic/bandwidth)
Network performance can be divided into
three simple parts that explain network
Rule of Thumb: Transport time per
performance impacts. These parts include
display = Mbpd / minimum gateway
network transport time (similar to software
bandwidth (Mbps)
service time on hardware platforms),
network travel time (this is the round-trip
packet travel time we call latency) times the Sample calculation:
number of round-trips (we call chatter), and
queue time (wait time for a transmission to • Enterprise geodatabase traffic = 10
get on the network due to network Mbpd
congestion). • Minimum bandwidth connection = 10
Mbps
• Network transport time = 1 sec (10
What is network transport time? Mbpd/10 Mbps)
Network transport time examples
Simple match can explain many of the

performance impacts we see from limited
bandwidth connections. Many people do not
bother to do the math, so here we try to
identify the value of doing the homework.
Most network design problems are quite
simple to avoid if you do your homework.
Figure 6.10 Network transport time is the time The data transfer traffic and the network
required to process data through the network bandwidth can be used to estimate
connection. minimum network transport times for a
single map display transaction.
Figure 6.10 shows a simple overview of a
WAN connection. Network transport time Rule of Thumb: A typical medium
is the time required for the network gateway complexity GIS application requires up
(switch, router, modem) to process the to 2 MB of data to generate a new map
traffic on the network. The network display. Typical terminal/browser display
connection design constraint is established traffic is about 200 KB of data for the
by the weakest (lowest bandwidth) shared same medium complexity map display.
network component within the transmission
route.

Figure 6.11 Typical map display transport time.
Figure 6.11 shows typical data transfer • When selecting data from a shapefile,
requirements in megabytes, shows the the total file (file index for large spatial
conversion to megabits of traffic for data file formats) must be delivered to
transmission, includes any adjustments of the client for query processing.
this data performed by the communication • Data not required by the application is
protocol, and identifies the total volume of rejected at the client location.
• This accounts for the considerable
traffic in megabits that must be transmitted
amount of traffic overhead experienced
(protocol overhead may be greater than
by these communications.
what was used in this sample conversion).
ArcGIS direct connect client/server
The minimum data transport times are configurations support DBMS query
calculated for five standard bandwidth processing on the server platform.
solutions.
• The query processing includes locating
• 1.54 megabits per second (Mbps) and 6 the requested data and filtering that data
Mbps for typical WAN communications so only the specific data extent
• 10 Mbps, 100 Mbps, and 1 gigabit per requested by the client is returned over
second (Gbps) for LAN communications the network.
• If the client application requests a small
Warning: Any existing data traffic on volume of data (e.g., point data or a
shared network segments would single parcel in a parcel layer), the
increase these network transfer times. resulting data transfer can be small and
network transport time would improve
File server configurations support query accordingly.
from the client applications.

Best practice: Desktop applications using either our geodatabase replication
accessing file or DBMS data sources services or disconnected editing using
perform best in a local area network feature services.
environment. Transaction-based
services, or persistent Windows Best practice: Citrix XenApp and Virtual
Terminal Server connections, provide Desktop Infrastructure (VDI)
reliable environments for processing architectures provide the best design
over less stable wide area network solutions for remote client ArcGIS for
connections. Desktop operations with a centralized
Warning: Several problems can occur DBMS maintenance environment located
when connecting a client application to a across a WAN.
DBMS across a WAN connection.
DBMS connection across a WAN What is network latency?
A DBMS connection across a WAN

connection can result in the following:
• Increased traffic between the client

application and the DBMS is only one
consideration.
• Network latency for chatty
communications contributes to
significant performance degradation and
is addressed in the following section. Figure 6.12 Network latency is the time is the round-
• A third issue is the need for a persistent trip travel time for a single communication packet.
connection between the client
application and the DBMS. Figure 6.12 shows how latency impacts
o If the client network connection display performance. Network latency can
is lost during a DBMS sneak up on you. It does not exist in the
transaction the data may not be local network (if you have good switches
received by the DBMS. and routers), and may not show up in your
o When the network connection is testing. It is not easy to identify how chatty
restored, the DBMS and client your software application is. This can be a
application session may not be big problem for remote users — Satellite
in sync. connections, for example, can have latency
o Once the application session is exceeding 1 second for each round trip.
out of sync, the DBMS will Wireless latency values can also be high.
protect data integrity and roll
back to the last database Network latency is the round-trip travel time
commit. for a single communication packet. Many
o This can result in lost work when applications require several sequential
accessing a DBMS over an round trips to the server and back to
unstable network connection – complete the display transaction. Each
an event that can contribute to round trip is a chat, and most applications
user frustration and reduced require several sequential round trips to the
productivity. server to complete a display transaction.
The chattier the application, the more
When you must communicate with a DBMS latency will impact display performance.
over a WAN, consider distributing the data

Total latency delay = latency x chatter transactions are required to support a single
map display.
• Latency = measured round-trip
travel time (milliseconds) Network latency can impact bandwidth
• Workflow chatter = number of utilization over long WAN distances.
sequential round trips for an average Latency is easy to measure, using a simple
display Windows command prompt (tracert "server
host name"). Results provide the number of
Rule of Thumb used for design network hops and the associated network
purposes: latency time for a single trip.
• Medium-complexity Enterprise/File The network latency example compares

Geodatabase chatter = 200 latency impact for local clients (1 msec
• Web/terminal client chatter = 10 latency) with remote clients (30 msec
latency).
Network latency example
For LAN environments, network latency is
very low (typically < 1 millisecond per round
trip to the server).
• Many trips between the server and client

have limited performance impact.
• The primary system latency contribution
is client and server processing service
times.
For longer WAN distances that involve

several router hops, there can be a
measurable network latency delay. For
chatty database connections, network
Figure 6.13 Latency impact for local clients is minimal. latency can have a measurable
Performance impacts for remote clients with
measurable levels of latency can cause noticeable
performance impact.
delays with chatty database communications.
• Total transaction time over the WAN
Figure 6.13 shows a simple example (including cumulative network latency) is
demonstrating latency performance 9.65 seconds.
impacts. Several similar use-cases have • The maximum bandwidth utilized by a
played out in system design consulting and single user on this WAN connection is
performance tuning. A simple ping 2.1 Mbps.
command can identify a potential latency
performance issue for chatty applications. Best practice: Latency should be
reviewed for performance impacts when
Normal database access protocols are remote clients are accessing chatty
"chatty," which means a typical database applications over long distances or
query requires a large number of trips to unstable communication connections.
and from the server to complete the client
display transaction. There are several trips Currently, many global WAN connections
to and from the server (query transactions) include satellite communication links. The
for each layer in the map display. Following fastest communication transfer is limited by
examples assume that 200 query the speed of light, which for very long

distances will introduce a minimum
bandwidth latency that technology will not
overcome. Good performance over WAN
environments results from protocols that
minimize trips (communication chatter)
between the client and server platforms.
Warning: Latency for en-route satellite

communication hops can be over 1
second, generating unacceptable delays
for chatty applications.
CPT network latency performance delays Figure 6.14 Shared network with no collisions is
shown on the top diagram. Shared network with
Network chatter is included in each random collisions is shown on the bottom graphic.
workflow on the CPT Workflow tab. Network
latency is defined on the remote site Figure 6.14 shows multiple user
network segments for each user workflow. communication packet transmissions on
The CPT will calculate network latency shared network segments. The top diagram
delays by multiplying network latency by shows random network traffic transmissions
workflow chatter and include the result in with no packet collisions, and the bottom
calculating display response time. diagram shows increased random traffic
transmissions with a packet collision. The
CPT Calculator tab total number of clients that can share a
single network segment will vary with
Latency is addressed for each remote network traffic transport time (size of data
location in the CPT Calculator tab. transfer and network bandwidth). Only one
client data frame can be transmitted over a
shared network segment at any time, and
CPT Design tab
display response time will slow down when
user communications are transmitted at the
Latency is addressed for each remote site same time due to packet collisions (network
network in the CPT Design tab. contention).
Shared network With older switch technology, multiple

transmissions at the same time on a shared
performance Ethernet network segment would result in
collisions. When a collision occurs, each
Network contention is all about busy client would delay and then retransmit to
networks. Network contention delays are complete data frame delivery. Ethernet
similar to what we experience in rush-hour segments would fail rapidly during
traffic on a freeway. It is really important to saturation because of the high number of
establish sufficient network capacity for transmissions.
peak traffic loads — this is often the time
that you need to get work done, and if you
Best practice: Maintain peak traffic for
do not have sufficient bandwidth the system
shared segments to levels below 25–35
will fail to support your critical productivity
percent utilization with older Ethernet
needs.
hub technology.

Ethernet switches today include options for ArcGIS workflow server output formats:
configuring shared segments in a full duplex
mode, which, configured properly, take • JPEG is recommended for imagery
advantage of switch buffer cache and output.
improve transmission efficiency. • PNG8 is a light-weight Portable Network
Graphics lossless data compression
Best practice: Plan peak traffic for supporting up to 256 colors and
shared segments to levels below 60 transparency. Can be used for business
percent utilization with full duplex layers where high color contrast is not
Ethernet switches. required.
• PNG24 is a medium-weight Portable
Network traffic adjustments Network Graphics lossless data
compression supporting up to 16 million
Network planning factors provide a starting colors and transparency overlay.
point for completing our network suitability Provides more color depth and
analysis. A baseline estimate of the network transparency control than PNG8.
traffic per display for each user workflow is • PNG32 is the highest quality Portable
established from software technology Network Graphics lossless data
pattern templates and internal benchmark compression. Provides the best support
testing. The actual traffic per display for a for graphics with gradients, varying
specific user case will also depend on the colors, rounded edges, and
client communication protocol (server output transparency.
format) and the workflow data source. • PDF is Adobe's Portable Document
Format. Delivers high-quality plot files
from your map document.
• Feature service delivers GIS feature
Output format
data to the client for editing purposes.
• ICA is the Independent Computing
Architecture protocol developed by Citrix
for communications between Windows
Terminal Servers and Citrix remote
client terminal displays.
Note: Output performance adjustment

factors are included in a look-up table
maintained on the CPT Calculator tab.
Data source format)

Figure 6.15 Network traffic adjustments for various
ArcGIS for Server output formats.
A series of benchmark tests were performed
A series of benchmark tests were performed to evaluate traffic impacts when using
to evaluate traffic impacts between available different data source formats. Results of
ArcGIS for Server map output formats. these benchmarks were used to establish
Results of these benchmarks were used to baseline adjustment factors for use in
establish baseline adjustment factors for generating custom data flow estimates for
use in generating custom data flow variable workflow data source formats.
estimates for variable server output formats.
Figure 6.15 shows the traffic adjustment
factors applied to the baseline client display
traffic based on selected output format.

Best practice: Move data from a
shapefile to a file geodatabase for best
performance.
Figure 6.16 Network traffic adjustments for different

GIS map data source formats.
Figure 6.17 Network traffic adjustments for different

Figure 6.16 shows the traffic adjustment imagery data source formats.
traffic based on selected GIS mapping
Figure 6.17 shows the traffic adjustment
vector data source formats.
traffic based on selected GIS imagery data
Values show relative performance source formats
compared with an enterprise geodatabase
data source.
Values show relative performance
compared with Imagery served from an
• DB DBMS: Baseline enterprise ArcSDE geodatabase data source.
geodatabase data source
• Small file geodatabase: Small file
• EGD: Imagery stored in an Enterprise
geodatabase data source
geodatabase.
• TIFF_uncompressed: Uncompressed
Best practice: A small file geodatabase Tag Image File Format; uncompressed
works well for hosted web services. It is files can be quite large.
recommended that a copy of the data be • TIFFLZW_compress: LZW is a lossless
deployed on the local disk for each compression. Works best with raster
ArcGIS for Server machine. data that has solid color ranges
available for compression. Less
• Large file geodatabase: Large file effective with imagery where there is
geodatabase data source color variation between pixels.
• Small shapefile: Small shapefile data • TIFFJPG_compress: Lossy JPEG
source (data source fits in application compression of a TIFF file.
server memory) • JPG2000: Upgraded JPEG (Joint
• Medium shapefile: Medium shapefile Photographic Experts Group)
data source (medium-size county compression algorithms published in
dataset) 2000. Supports lossless and lossy
• Large shapefile: Performance with a compressions. Published as an ISO
large shapefile data source can be standard. Provides several
marginal enhancements over standard JPEG
compression. Requires

encoders/decoders that are complex Note: Data source format performance
and computationally demanding. adjustment factors are included in a look-up
• MRSID: LizardTech Multi-Resolution table maintained on the CPT Calculator tab.
Seamless Image Database format that CPT Calculator workflow network
uses a mix of techniques supporting performance analysis
lossless and lossy compressed datasets
for use by the GIS community.
A single workflow network performance
• ECW: Enhanced Compression Wavelet
analysis can be completed using the CPT
format optimized for aerial and satellite
Calculator.
imagery. Developed by Earth Resource
Mapping, and is now owned by ERDAS,
which is owned by Intergraph. Best practice: CPT Calculator is a useful
• IMG_ERDAS: ERDAS Imagine image tool for evaluating single workflow
file format. network traffic contributions and the
impact of available network capacity on
remote site user productivity.
A variety of imagery sources and formats
are available, and methods for collecting Warning: The CPT Calculator is limited
and sharing imagery data sources is of high to addressing a single workflow
interest within the GIS community. environment. Additional network traffic
Technology is rapidly evolving to capture from other business operations will need
and deliver opportunities for leveraging and to be included to complete the analysis.
sharing imagery data products.
Enterprise network architecture

GIS user workflows access applications,
services, and/or data sources maintained
within the enterprise data center. Users may
be located on the LAN, at remote locations
over the WAN, or operate from mobile
clients over wireless WAN or Internet
connections. Remote and local users may
also access a variety of web services over
the public Internet.
The enterprise GIS network communication

Figure 6.18 Enterprise network component diagram shows GIS users located at three
architecture provides the infrastructure for connecting remote sites: one connected over an
business operations. Internet connection and the other two
connected over a WAN connection.
Figure 6.18 shows a typical IT drawing, and Additional GIS users are located within the
the colors used match the network locations central LAN environment.
used in the CPT. The CPT Design tab is an
information product matching the IT • Remote user site connections are
diagram, once you understand the network shown in green.
and user workflow relationships. • Central data center connections are
shown in gray.

• Data resources and web services are
published from the central data center.
• Remote site users access central data
center resources across the Internet and
WAN connections.
Network connections within the local LAN

infrastructure can normally be adjusted to
accommodate peak traffic loads without Figure 6.19 Results of a user workflow analysis
significant cost impacts. Remote WAN and review. Establishes requirements for the system
Internet connections often present architecture design.
constraints that must be addressed in the
system design, and may limit the viable Results of a user workflow analysis review
software technology deployment options. establishes performance parameters for the
system architecture design. Figure 6.19
Warning: Network bandwidth limitations shows a sample of the results of a workflow
(not enough bandwidth to handle peak analysis review (Business Requirements
traffic flow) can seriously constrict user Summary):
performance during peak operational
loads. • User requirements include three
workflows (DeskEdit, DeskView, and
CPT Design user requirements module WebMap)
configuration establishes a foundation for • Users are located at different network
completing the network suitability analysis. locations (local users, remote site 1,
remote site 2, and public internet)
Best practice: GIS user requirements • Peak workflow usage is identified for
analysis should be completed before each network location
establishing the final system o Local Area Network (peak users
architecture design. for each workflow)
o Remote site 1 (peak users for
The CPT Design requirements analysis each workflow)
module layout makes it possible to o Remote site 2 (peak users for
represent workflow display traffic flow each workflow)
across the appropriate service provider o Public Internet connection (peak
network connections. transactions per hour for
WebMap services)
Best practice: System design
configuration begins with results from a CPT Design user requirements workflow
user workflow analysis. loads analysis
CPT Design requirements section is

configured to represent the user locations
and network connectivity. Once the
requirements section is properly configured,
the CPT (Excel) completes the network
suitability analysis.

Best practice: Configuring the administrator and included in the
requirements analysis module network infrastructure upgrade budget.
establishes a foundation for completing
the CPT Design tab network suitability
analysis.
Network Suitability Analysis
Many network administrators establish and
Network pricing model
maintain metrics on network utilization,
which help them estimate increased
network demands when planning for future
user deployments. The CPT Design tab
includes a network suitability analysis which
uses display traffic from the Workflow tab to
compute peak site traffic workflow loads.
Sum of all site traffic workflows is compared
with the site bandwidth to identify peak
Figure 6.20 Network pricing model for Wide Area
network utilization. Network bandwidth
Network (WAN) and Internet service provider rates. should be configured at about twice the
peak traffic loads to avoid contention and
Network bandwidth can be a critical performance impacts.
consideration for a successful system
architecture design, with significant impacts
on system performance, overall deployment Computing network utilization
cost, and optimum design solution. Figure
6.20 shares a network service provider cost
model used within the System Design
Strategies capacity planning
demonstrations.
Network service provider costs should be

considered when evaluating Enterprise level
GIS design solutions.
• Network bandwidth has an impact on

remote user productivity with every GIS
design. Figure 6.21 User location and peak workflow loads
• Technology change (i.e., higher can be used to identify site traffic. Traffic utilization is
resolution clients) increase GIS site traffic / site bandwidth.
workflow traffic requirements.
• Dynamic feature services generate Figure 6.21 shows the user workflow and
more traffic than standard mapping peak site traffic. The CPT Design tab
services. includes a network suitability analysis which
uses display traffic from the Workflow tab to
Warning: Network service rates can be a compute peak site traffic workflow loads.
significant performance/cost
consideration for an optimum system • Identify user workflows. These are the
design. workflows that will generate the network
traffic.
Best practice: Network bandwidth • Identify average display traffic (Mbpd).
upgrade recommendations must be This is the traffic generated across the
coordinated with the network

network for each user display
transaction. Network traffic queuing
• Identify user locations. Identify the
network route between the user location
and the data center.
• Compute workflow traffic. Workflow
traffic is the product of user productivity
and display traffic.
o Traffic = productivity (DPM) x
Mbpd / 60 seconds
• Compute site traffic. Total traffic that will
be communicated over the shared
network connections.
Figure 6.22 Top diagram shows traffic peaks shown
o Site Traffic = Sum (workflow below maximum network capacity. Lower diagram
traffic for that site) shows traffic peaks above maximum network
capacity. User productivity is reduced due to heavy
Productivity is a measure of user display traffic loads.
activity represented in average displays per
minute. Each new user display requires a Models use average loads. Live loads have
workflow transaction that generates network peaks and valleys, which are quite
traffic. The total amount of traffic divided by predictable using queuing theory (because
available bandwidth is the network you are dealing with very large random
utilization. The sum of all the site traffic populations). If the network gets too busy,
workflows is compared with the site user productivity will slow down. If you are
bandwidth to identify peak network way off in your network design, work can
utilization. slow to a crawl. Network suitability analysis
should be done for every design to avoid
Calculating network utilization. performance failures.
• Identify site bandwidth. Shared networks must adjust traffic flow to

• Identify peak site traffic. accommodate random transmission arrival
• Compute network utilization. times. When peak traffic loads remain within
o Utilization = traffic / bandwidth the available bandwidth, users will not
experience network delays. On the other
Note: The network suitability analysis hand, when concurrent transaction arrival
evaluates how much of the available times exceed available network capacity,
bandwidth (utilization) will be required to users will experience delays. Concurrent
meet peak system loads (if you have too transactions must wait for network
much traffic for the available bandwidth). connection, resulting in transmission delays.
• Delays increase during heavy traffic

loads as the network reaches saturation.
• Wait time (queue time) is a function of
network transport time and bandwidth
utilization.
• Higher traffic per display and busier
shared network segments contribute to
longer transmission delays.

User productivity It is important in a design analysis to
validate user productivity. If established use
For many GIS workflows, the user computer productivity can not be maintained due to
display response time can be a factor in system contention, the estimated system
reducing user productivity. Slow display loads would be more than what is possible
response times reduce user productivity. in a real environment.
Validating user productivity.
• Cycle time = 60 seconds / user

productivity
• Computed think time = Cycle time -
workflow response time
• Minimum think time is the minimum time
required for user decision and input
(defined by workflow).
Figure 6.23 Network response time = Network
transport time + Network queue time. Computed think The workflow is valid if computed think time
time = user productivity cycle time - workflow > minimum think time. User productivity
response time. The Workflow Performance Summary must be adjusted if the workflow response
provides workflow response time as a stack of service
times, queue times, and latency delays. time does not support minimum think time.
Figure 6.23 shows the factors that make up Warning: The workflow is not valid if
user display response time. The total minimum think time exceeds computed
service time includes the computer think time.
processing and network transport times for
each display. The queue time includes any Figure 6.23 shows the CPT Design
system wait times due to processing Workflow Performance Summary. The
contention. The total response time also Workflow Performance Summary shows
includes network latency delays. Queue workflow response time as a stack of
time increases as the network and platform software service times, network transport
system components get busy, and the times, associated queue times, and network
queue time profile is predictable for large latency (available once system level
populations of random transmission performance validation and any required
requests (computer processing workflow workflow productivity adjustments are
transactions). [Queue time calculations] will complete).
be discussed in more detail in chapter 10.
Workflow performance is the sum of the
Display response times can be computed following component parameters:
for each workflow, based on the total of all
system component service times and queue • Software service times
times during peak system loads. • Network transport time
• Associated component queue times
• Calculate network queue times. • Network latency delays
• Calculate network transport times.
• Network response time = transport +
queue times + latency.
• Calculate workflow display response
times (platform response times +
network response times).

User productivity adjustment Warning: If the display response time
plus the minimum think time exceeds six
Normally, once you identify that you do not seconds, the user will not be able to
have enough bandwidth, you talk with the maintain a productivity of 10 displays a
network administrator about network minute and will have to slow down.
upgrades. If you are not able to upgrade the
network, you can look for another • The CPT Design tab calculates display
distributed GIS solution pattern that will response time and available user think
avoid the network bottlenecks. The RESET time for each user workflow.
ADJUST function is useful in identifying how • If the computed think time is less than
wrong a solution might be - what are the the workflow minimum think time, the
user productivity impacts? This is often workflow productivity must be reduced
helpful in validating a more costly end to show a valid workflow load.
solution.
The CPT Design tab has an ADJUST
function, cell AF2, that automatically adjusts
user productivity to honor identified
minimum think time requirements.
Best practice: The RESET ADJUST

function can be used to identify
maximum user productivity available
with increasing bandwidth.
Warning: The RESET ADJUST function
Figure 6.24 Workflow is invalid when minimum think does not perform well with very small
time is greater than computed think time. The CPT network transport times. Network
Design tab RESET ADJUST function will reduce user bandwidth should be increased to
productivity until calculated think time equals
handle traffic flow.
minimum think time. Workflow is valid when minimum
think time is not greater than computed think time.
Figure 6.24 shows a valid workflow
Figure 6.24 shows what happens during following the use productivity adjustment.
heavy traffic conditions, when network Once minimum think time equals computed
bandwidth is saturated beyond capacity think time, the CPT identifies a valid
limits. workflow.
• User productivity determines the • Reduced workflow productivity

average server transaction request rate. increases workflow cycle time.
o 10 displays per minute, or one • Reduced workflow productivity reduces
display every six seconds. network utilization.
• Users will start to slow down when they • Reduced network utilization reduces
are not able to maintain their network contention (queue time).
productivity rate. • Reduced queue time with increased
• To maintain productivity, the user must cycle time increases computed user
receive the requested map display from think time.
the server (response time) and have
sufficient time (minimum think time) to Note: There will be further discussion about
review the display information and system performance factors and
submit the next transaction request. relationships that impact system capacity
and user productivity in Lesson 10.

Best practice: Network suitability Best practice: Increase network
analysis is an important part of system bandwidth to roughly twice expected
architecture design; making sure you peak traffic flows.
have sufficient network infrastructure to
ensure remote user productivity is
critical for implementation success. RESET ADJUST function
CPT Design network suitability The RESET ADJUST function on the CPT
analysis Design tab can be used to demonstrate
impacts on user productivity due to network
Initial user requirements loads analysis bandwidth constraints. The RESET
ADJUST function reduces user productivity
Network Suitability Analysis is completed by for all workflows adjusting system loads to
the CPT Design tab as you configure user fit within existing infrastructure capacity
locations, input peak workflow loads, and constraints.
identify network bandwidth.
Workflow performance summary
Figure 6.25 The Workflow Performance Summary shows performance once you have a valid user workflow.
Figure 6.25 shows the Workflow The Workflow Performance Summary will
Performance Summary following the identify queue times for the reduced
productivity adjustment. Once you have productivity workflows.
resolved all workflows to a valid workflow Validated design solution
solution, you can review the workflow
performance summary to evaluate workflow Once you complete your network suitability
response times for each user site location. analysis, you should work with the network
administrator to identify appropriate network
bandwidth upgrades.

Best practice: Network bandwidth Network performance parameter look-up list
should be roughly twice the expected
peak traffic loads to avoid performance Network performance parameters are
bottlenecks. identified on the CPT Workflow tab.
Warning: Network traffic utilization over Workflow display chatter, client display
60 percent could contribute to network traffic, and database traffic columns are
contention. established when creating a user workflow
and are available in a look-up table located
Once you agree on bandwidth upgrades, on the CPT Workflow tab.
you can enter them in the CPT Design to
complete your analysis.
Network contribution to Web performance
Figure 6.26 Network bandwidth connectivity is a primary factor in slow Web client display performance.
Figure 6.26 shows the network contribution • Network performance improves with
to Web display performance. Network increased bandwidth.
performance counts. Network transport time o 45 Mbps bandwidth connection
is a major contribution to web client display is 30 times faster than a 1.5
response times. Mbps connection.
• Light display traffic performs much
• Web services tend to deliver higher better than heavy display traffic.
display traffic than WTS ArcGIS for
Desktop clients. Best practice: Network performance
• Server display render time can be much impacts should be considered carefully
less a factor than the network during design and deployment of web
connection. services.

Conclusion
Network bandwidth suitability is an important part of the system design process. Expanding GIS
technology patterns, particularly over wide area and mobile cellular networks, have place heavy
demands on the network infrastructure. Available network capacity, particularly the last mile in
connecting to the client, can determine if a particular software technology pattern is suited for
your business solution. It is best to identify network contention limitations during the design
process. Too often we see clients deploy solutions for client locations with insufficient network
infrastructure. These mistakes can be avoided by completing a network suitability analysis as
part of your system architecture design.
The following video provides an overview of the CPT Calculator network traffic functions. The
video also provides an overview of the CPT Design User Requirements Module and shows how
the CPT Design completes the network suitability analysis.
CPT Capacity Planning videos

Platform Performance
Fall 2017 Platform Performance 41st Edition
Chapter 3 (Software Performance)

discussed some best practices for
publishing high performance map services,
and the importance of selecting the right
software technology to support your
business needs. This chapter will focus on
hardware platform performance, and share
the value of selecting the right computer
technology to support your system
performance needs.
Selecting the right hardware will improve

user performance, reduce overall system
cost, and establish a foundation for building
effective GIS operations. Selecting the
wrong hardware can contribute to
implementation failure - spending money on Figure 7.1 User Performance Expectations
a system that will not support your business
needs. To develop a system design, it is necessary
to identify user performance needs. User
Hardware vendors do not know what productivity requirements can be
hardware is required to satisfy your GIS represented by workstation platforms
needs. This chapter shares the system selected by users to support their computing
architecture design methodology developed needs. GIS users have never been satisfied
to help you select the right hardware for with platform performance, and each year
your planned GIS operations. This chapter power users are looking for the best
also shares information for justifying computer technology available to support
hardware purchases based on expected their processing needs.
return on investment.
As platform technology continues to
Platform Performance improve, user performance expectations
may change. It is not clear just when
Baseline computers will be fast enough for GIS
professionals – there is always more we
The world we live in today is experiencing could do if we just had the power. As user
the benefits of rapid technology change. productivity improves, application and data
Technology advancements are directly servers must be upgraded to service the
impacting GIS user productivity—the way increasing user desktop processing
we all deal with information and contribute requirements.
to our environment. Our ability to manage
and take advantage of technology benefits GIS user performance expectations have
can contribute to our success in business changed dramatically over the past 10
and in our personal life. years. This change in user productivity is
caused primarily by faster platform
performance and lower hardware costs.
Figure 7.1 identifies the favorite hardware
Platform Performance 7.1

platforms selected by GIS users over the Xeon E3-2637v4 4-core (1 chip) 3500 MHz
past 10 years. High performance desktop platform
workstations have made a primary
contribution to improving user productivity • includes four of the fastest processors
and expanding GIS technology capabilities. released by Intel in early 2016.
• ArcGIS Desktop workstation
Each year we review hardware vendor recommended memory is increased to
technology to identify the best available 16 GB to accommodate expanding GIS
platform for GIS professional workstation use of large imagery files, concurrent
users. The highest performing platform is use with ArcGIS Pro, and increasing
used to identify our performance baseline emphasis on time aware geoprocessing
for each calendar year. The Xeon E3- analysis.
2637v4 4-core (1 chip) 3500 MHz was
identified as our favorite 2016 and 2017
workstation.
Performance Baseline history
2017 Arc17 Performance Baseline = 64 (SPECrate_int2006 per core baseline)
Figure 7.2 Platform Performance Baseline
Figure 7.2 provides a graphic overview of platform performance improvements

relative Intel processor performance over contribute to improved business productivity
the past 10 years. Platform per core and system computer capacity, reducing the
performance today is over 4 times faster overall cost of automated business systems.
than 10 years ago. Hardware vendor

The boxes at the bottom of the chart available platform technology. These
represent the performance baselines used performance baselines are reviewed and
to support the Esri capacity planning models updated each year to keep pace with the
over the past 10 years. Performance rapidly changing hardware technology.
baselines are specified by calendar year
based on the per core performance of
Moore's Law
Figure 7.3 Moore’s Law performance history
The Intel founder Gordon E. Moore released continue to increase. The increase in
a paper in 1965 that predicted that the performance has been slower than in the
number of components in integrated circuits past, and more like a strait line performance
would double every year through at least over the past 10 years.
1975. His prediction, known today as
Moore’s law, has proven true over the last There is some discussion over the past
60 years and continues to contribute to couple of years questioning whether
computer performance gains. platform per core performance will continue
to improve as it has over the past 10 years.
Reducing the distance between integrated Moore's law deals with components getting
circuits has a direct impact on platform smaller and closer together on integrated
processor compute performance. A function circuits with each production design cycle.
that doubles every two years when plotted Distance between components today are
on a chart would produce an exponential approaching atomic size, and maybe there
growth curve. Figure 7.3 shows a plot of the are physical limits to how much faster the
Esri performance baselines with an chip will perform. Other factors such as
exponential curve overlay. From this chart, temperature and cooling limitations could
you can see that the performance gains we limit further processor performance gains,
have experienced over the past 10 years and performance per core could start to

level off with future chip releases virtual server deployment, while at the same
(represented by the red curve on the chart). time generate more heat and limit peak
High capacity chip configurations (16-core processing speeds (MHz); another factor
per chip, 24-core per chip, etc) are better for limiting per core performance gains.
Faster platforms provide more service with less hardware
Figure 7.4 Web mapping service peak concurrent users and peak service throughput for entry level software license
with medium complexity mapping services is plotted relative to platform performance increases over the past 10
years. The general trend drives reduced cost for GIS software deployment.
Figure 7.4 represents the relationship Web deployment timelines are significantly
between server platform performance and reduced with new software, reducing
peak entry level Web mapping service software development and deployment
throughput. The chart shows rapidly expenses. Web mapping software services
increasing software license service capacity that took over 6 months to develop and
rendered by platform performance deploy in 2007 can be deployed within
improvements shown in Figure 7.2. minutes to hours with 2017 technology.
Platform performance improvements reduce Improved hardware platform performance is

software cost. Dynamic web mapping driving a significant reduction in overall
services deployed using an entry level Enterprise GIS system cost.
ArcGIS Server software license in 2007
could support up to 40 concurrent users
(14,473 TPH). Those same mapping
services deployed by an entry level ArcGIS
Server license with 2017 platform
technology can support over 285 concurrent
users (104,000 TPH) with more quality and
functionality.

Relative platform performance hardware vendors initially established in the
late 1980s for the purpose of establishing
guidelines for conducting and sharing
relative platform performance measures.
They have developed a variety of standard
benchmarks with testing performed by the
hardware vendors and published on the
SPEC Web site. Additional information on
Intel processor configurations is available
on the Intel Architecture site.
Figure 7.5 How do we Handle Platform Performance
Change? The SPEC compute-intensive benchmarks
have been used by Esri as a reference for
Knowing how to account for platform relative platform performance since 1992.
technology change is fundamental to The system architecture design platform
understanding capacity planning. Figure 7.5 sizing models used in conjunction with
identifies a simple relationship that we have these relative performance measurements
used since 1992 to relate platform have supported Esri customer capacity
performance with capacity planning. planning since that time. The SPEC
benchmarks were updated in 1996, 2000,
The relationship simply states that if one and 2006 to accommodate technology
can determine the amount of work (peak changes and improve relative performance
throughput) that can be supported by server measurements.
A and identify the relative peak throughput
performance between server A and server SPEC provides a separate set of integer
B, then one can estimate the amount of and floating point CPU intensive
work that can be supported by server B. benchmarks. Computer processor core can
This relationship is true for single-core and be optimized to support integer or floating
for multi-core servers. point calculations, and performance can be
very different between these environments.
Platform capacity test results with ArcGIS
Platform performance resources Desktop and Server software have tracked
quite close to the SPEC CPU integer
Having a fair measure for relative platform relative performance throughput (rate)
performance and capacity is important. benchmark results. This confirms that the
Selection of an appropriate performance ArcGIS component software code
benchmark and agreement on how the predominantly uses integer calculations.
testing will be accomplished and published The SPEC CPU integer benchmarks
are all very sensitive hardware marketing provide the best relative platform
issues. You need a performance measure performance estimates for representing
that is accepted by the vendors – preferably ArcGIS software technology.
one they provide themselves.
SPEC provides two methods for conducting
SPEC performance benchmarks and publishing their CPU integer benchmark
results. The SPECint2006 is a speed
SPEC CPU integer relative performance benchmark measuring execution time for a
throughput (rate) benchmark results single benchmark instance and then uses
the result for calculating relative platform
The Standard Performance Evaluation performance. The SPECint_rate2006 is a
Corporation (SPEC) is a consortium of throughput benchmark, and is conducted

using several concurrent benchmark baseline values used as a lookup table by
instances (one for each core thread) and the CPT Calculator, Design, Test, and
measures executable instance cycles over a Favorites tabs.
24-hour period. The SPECint_rate2006
benchmark results are used for relative Published vendor benchmark values are
platform capacity planning metrics in the used to identify relative throughput and
Esri system architecture design capacity performance for selected hardware
planning models. platforms. Platforms are arranged by vendor
and year in two lookup lists. Desktop
There are two results published on the candidates are located at the top of the list.
SPEC site for each benchmark, the Server candidates are located at the bottom
conservative (baseline) and the aggressive of the list. Project platform candidates are
(result) values. The conservative baseline located in the middle of the list and included
values are published first by the vendors, with the Desktop and Server list selections.
and the aggressive values are published
later following additional tuning efforts. The Capacity Planning Tool updates page is
Either values can be used to estimate the primary source for the platform
relative server performance, although the performance metrics. Information from the
conservative benchmarks provide the most SPEC Web site is entered into the CPT
conservative relative performance estimate Hardware tab for capacity planning. Copy of
(removes tuning sensitivities). We the SPEC benchmark information is
recommend using the conservative provided in a HardwareSPEC Excel
benchmarks for capacity planning, and the workbook for easy access. The SPEC
published SPEC CPU benchmark baseline benchmark values are used to adjust
values are used in the Capacity Planning baseline service times to selected platform
tool. service times for capacity planning analysis.
Several benchmarks are published on the HardwareSPEC Excel Workbook

SPEC Web site. You will need to select and
go to the SPECrate2006 Rates and then The Capacity Planning Tool updates site
scroll down to the configurable request shares a HardwareSPEC workbook with an
selection - you can then select specific Excel table of platform relative performance
items that you want included in your display values from the published
query. I like to include the processor and the SPECrate_integer benchmarks.
Processor MHz in my display, which was
not included in the default selection. The Adding a new platform to the CPT Hardware
Processor Characteristics include maximum tab
Turbo boost MHz which can be used to
estimate maximum performance at low New hardware platform benchmark values
utilization levels. are published on the SPEC Web site each
month throughout the year, so the platform
More information on the SPEC CPU you need for your design analysis may not
benchmarks can be found on the SPEC be included in your version of the CPT. You
Web site. can locate the new benchmark values on
the SPEC Web site and then add them to
CPT Hardware tab your CPT Workflow tab.
CPT Hardware tab includes a list of Desktop

and Server SPEC CPU platform benchmark

Web map display performance history
Figure 7.6 Time to Produce a Map. Chart shows service time improvements for medium complexity mapping services
based on platform performance improvements shown in Figure 7.2.
The change in hardware performance over seconds) which we believe may open new
the years has introduced unique challenges opportunities for GIS analysis and display.
for capacity planning and for software Traditional heavy (5x Medium > 3x Heavy)
vendors trying to satisfy customer map displays can now be rendered in less
performance and scalability expectations. than 1 second, suggesting hardware
Understanding how to represent hardware technology may no longer be a limitation on
performance differences is critical when GIS user productivity. IT departments see
addressing capacity planning, performance, this as an opportunity to buy higher capacity
and scalability issues. platforms and leverage virtual server
environments and cloud computing to
Figure 7.6 shows how user expectations simplify their administration workload
have changed over the past 10 years. An (exchanging user display performance for
ArcGIS Desktop heavy dynamic map lower administration costs). I expect GIS
display processing time in CY2007 would users will see this as an opportunity to
take over 2 seconds. That same map incorporate more complex analysis into their
display today can be rendered in less than user workflows, leveraging more compute
0.3 seconds – over 7 times faster than just intensive statistical analysis, logistics
10 years earlier. Most of this performance routing functions, and business analytics for
gain can be accounted for by faster use in their standard business workflows.
processor core and a new ArcGIS Server Heavier processing workflows will require
display rendering engine. continued hardware performance
improvements to keep user productivity at a
Figure 7.6 shows a minimum user peak level.
performance expectation range (1-2

2017 Technology Changes
Figure 7.7 2017 Technology Changes
Figure 7.7 highlights the technology Data Center consolidation can save
changes that are making a difference in operational costs. Cloud hosting can
2017. We have higher capacity servers, change how we manage and support our
more processor cores per chip, and more enterprise business operations.
chips per server.
Warning: All of the Cloud vendor
Processor vendors are responding to new administrative savings may not be
data center platform environments. Virtual passed on to the customer.
server deployments are becoming standard
practice for enterprise data centers. Hardware processing performance this past
Vendors are introducing faster 4, 6, 8, 10, year was minimal. Delays in delivery of new
12, 14, 16, 18, and 22 core chips, with a 2017 processor cores have delayed
growing number of high capacity 4 chip performance improvements. 2017
servers. Data Center consolidation efforts performance baseline is the same as 2016.
are expanding rapidly. Cloud computing
solutions are being accepted as a viable Enterprise license agreements provide a
and cost-effective alternative to on-premise more adaptive and cost-effective way to
operations, and we are seeing an manage enterprise GIS operations.
expanding number of vendors with public
cloud offerings. The ArcGIS Enterprise platform deployed
online or on-premises with Portal for ArcGIS

introduces new opportunities for bringing Best Practice: ArcGIS Online and Portal
communities together. for ArcGIS provide rich platforms to
share, collaborate, and deploy your GIS
• Simplified licensing based on user roles. services within your organization or with
• New GIS Server roles increase Portal a broad community of users.
for ArcGIS capabilities.
• Portal collaboration supports distributed
systems of engagement.
Platform identification
Figure 7.8 Platform Identification
Hardware vendors provide a wide range of number of platform chip manufactures

choices at different performance levels for still competing in our marketplace (Intel
different user communities. New processors and AMD provide all of the Windows
may perform faster than older processors processor technology). These processor
that run at a higher clock speed (MHz), and chips are used in building all of the
processor speed is no longer a good hardware vendor platforms offerings.
measure of performance. Figure 7.8 shares • Total processor core. The total number
how vendors have responded to this of processor core identifies how many
problem, and the nomenclature we use to user requests can be processed at the
make sure we understand the platform we same time. Total core is a key
are talking about. parameter for establishing appropriate
memory and identifying the proper
CPT Platform terminology software configuration and platform
capacity. You may find vendors
• Chip processor number. Hardware identifying platforms by number of chips
vendors have identified specific model and how many cores per chip – you
numbers that are unique for each need to do your math to identify the total
processor chip configuration (E3- number of core. This can be confusing,
1720v5). Hardware vendors use these and for this reason the CPT terminology
chips as components in building their we use includes the total number of
server offerings. There are a limited processor core.

• Total chips per node. Total number of a socket, so the terms are used
chips included in the platform interchangeably. I use chip in the CPT
configuration is provided for information nomenclature because this is the term
purposes (not as important as used by SPEC and is shorter than using
understanding the total number of core). socket.
Some vendors refer to chips as sockets
– a chip is the board that holds the
integrated communication circuits and
the core processor. The chip plugs into
Platform Performance
Figure 7.9 Platform Performance Makes a Difference
Hardware vendor technology has been System processing capacity becomes very
changing over the past 5 years. Improved important. System availability and scalability
hardware performance has enabled are most important. The quality of the
deployment of a broad range of powerful information product (display and database
software and continues to improve user design) provided by the technology can
productivity. Sub-second server processing make a user's think time more productive.
times suggest that future user productivity Proper tradeoff between display quality and
gains will likely come from more loosely performance contributes to optimum user
coupled operations, higher capacity network productivity.
communications, disconnected processing,
mobile operations, pre-processed cached
maps, and more rapid access and
assimilation of distributed information
sources.

Hardware vendor performance gains • Intel Xeon E3-2637v3 4-core 2015 per-
core performance peaked at 57.4.
Much can be learned about server platform • Intel Xeon E5-1270v3 4-core 2014 per-
competition from vendor published SPEC core performance peaked at 53.
benchmarks. Figure 7.9 provides an • AMD Opteron's best 2012 per-core
overview of relative per core performance performance peaks just over 20. We
for key vendor-published benchmarks from have seen only minor AMD performance
2012 to 2017. gains since 2006.
• Most other processor vendors have
Intel processors are maintaining a strong dropped out of the competition.
performance leadership position over the
last 5 years. ArcGIS Server software is deployed on Intel
and AMD processors, supporting both Intel
• Intel Xeon E3-2637v4 4-core chips show Linux and Windows deployments. The Intel
2017 per-core performance at 59.9; platforms deliver a strong performance
expect some gains later in the year. IBM advantage over AMD, improving server
Power8 64-core chips show 2017 per- peak throughput capacity and reducing
core performance at 65.2, an impressive overall software licensing costs.
target for future Intel processor core
releases.
Processing speed drives platform throughput
Figure 7.10 Less processing time means more peak throughput
Figure 7.10 shows the relationship between processor core can service more requests.
display processing time and system More service requests per core means more
throughput. peak throughput.
There is an indirect relationship between ArcGIS Enterprise licensing is based on

server processing time and peak system number of hardware platform physical or
throughput. Faster per core processor virtual server cores (all core are treated
performance reduces service processing equal). More processor core implies more
time. Shorter processing time means each server throughput, with license cost based

on number of server core. Faster processor Warning: Deploying ArcGIS Server on
cores optimize platform peak throughput for platforms with slower processor core
a given server license (more bang for the will increase cost per service
buck). transaction.
2017 Intel processor performance
Figure 7.11 2017 Intel Processor Performance
The right platform selection is based on a • Ten core chip configurations (1, 2,
balance between server capacity, and 4 chip configurations)
performance, and power. Figure 7.11 • Twelve core chip configurations (1,
identifies the Intel processors available for 2, and 4 chip configurations)
the vendor 2017 hardware. • Fourteen core chip configurations (1,
2, and 4 chip configurations)
Intel provides ten 2017 platform commodity • Sixteen core chip configurations (1,
server deployment strategies. 2, and 4 chip configurations)
• Eighteen core chip configurations (1,
• Quad core chip configurations (1 2, and 4 chip configurations)
and 2 chip configurations) • Twenty two core chip configurations
• Six core chip configurations (1, 2, (1, 2, and 4 chip configurations)
and 4 chip configurations)
• Eight core chip configurations (1, 2, What platform should I buy? Quad and Six
and 4 chip configurations) core chip configurations provide the highest
per-core performance. Six and Eight core

chip configurations can support a higher quite impressive, making your final server
number of virtual servers per platform while decision a challenging choice (particularly
delivering excellent per core performance. for larger virtualized data center
Twelve, fourteen, sixteen, eighteen, and environments that can take advantage of
twenty-two core chip configurations can the higher capacity servers).
support a higher number of virtual servers
per platform at reduced per core Stay away from the slower performing
performance. platform models, they will likely end up
costing you more in software licensing.
The high core performance for the 2017 six,
eight, and ten core chip configurations is
2017 ArcGIS Server platform selection
Figure 7.12 Identifying the right Platform / How do we select the platform we want?
When you go to purchase a platform, system cost. The good news (for GIS users)
vendors are not very good at sharing their is that the best performing hardware often
performance numbers. I will say, to the delivers the lowest overall system cost. If
vendor’s credit, that they are good at you don’t do your homework, you might
providing their performance numbers on the miss the savings.
SPEC site (but not on their sales page). You
need to do your homework before you buy Figure 7.12 provides an overview of
your hardware. With GIS servers, platform platform configuration options provided by a
performance is important both for optimum vendor showing their relative per core
user productivity and to reduce overall performance and dollars per transaction.

The dollars per transaction is calculated by Establishing business needs.
dividing the total cost of the server divided
by the relative SPEC benchmark throughput Proper hardware selection depends on a
value. *Hardware pricing was based on list clear understanding of your business needs.
values. Platform pricing analysis will show how to
use the CPT to select the optimum platform
• Esri Software estimated at $10,000 per for publishing your mapping services.
core.
• SPEC baseline/core used for processor Completing the capacity planning
speed. analysis
• CPT Calculator used to evaluate virtual
server core requirements. Business workflow requirements are used to
identify the required server cores for each
CPT Best Buy Analysis analysis.
CPT can be used to complete a platform Best Buy:

best buy analysis. Sample platform best buy
analysis is provided in the Capacity • Xeon E5-2637v4 4-core platform
Planning Tool appendix. provides the best per-core performance.
• Xeon E5-2697V4 36-core platform has
CPT was designed to automate the system the lowest per-core platform cost.
architecture design analysis. It is particularly • Xeon E5-2667V4 16-core platform has
suited to translate user business the lowest per-core system cost when
requirements to appropriate hardware including GIS software.
platform selections.
ArcGIS Platform Sizing

The primary purpose for this Platform Performance chapter is to share best practices for
selecting the right vendor hardware platforms (platform sizing). Since 1992 our system design
consultants have helped customers identify proper hardware solutions. We found out early on
that it is not possible to have a successful GIS deployment if we don't get the hardware right.
For this reason, our system architecture design process was developed to translate business
needs to identified platform requirements.
Business needs are identified through a user workflow loads analysis. The system architecture
design analysis translates peak workflow loads to identified platform solutions. The CPT tools
can be used to complete the design analysis. This section will share how vendor platform
technology is changing, and show some rough estimates of what the available platform
configurations can deliver in an ArcGIS environment.

ArcGIS Desktop Platform Sizing
ArcGIS Desktop system requirements are identified in the ArcGIS 10.5 help. Figure 7.13 shows
Intel platform performance gains experienced over the past five years. The 2017 workstation
processors are over 40 percent faster than the processors that supported ArcGIS Desktop
workstation users in 2012. Additional memory and faster storage solutions can contribute to
additional performance gains. The advance of GIS technology is enriched by the remarkable
contributions provided by Esri's hardware partners.
Figure 7.13 Workstation Platform Recommendations
Workstation life cycle upgrades depend speed, not by the total number of
on user performance needs. available processor cores.
Note: Some ArcGIS geoprocessing
ArcGIS Desktop power user productivity is background services are multi-threaded,
often limited by processor per-core and can take advantage of additional
performance, and upgrading power user available processor core. ArcGIS Pro is
workstations can increase user productivity. multi-threaded and achieves optimum
Upgrade ArcGIS Desktop power user performance with hyper threading enabled
workstations whenever there is a large and at least 4 processor cores.
improvement in processor per-core
performance. Typical power user ArcGIS Desktop standard users normally
workstation life cycle is 1-2 years. work fine with slower display performance. It
is a good practice to upgrade ArcGIS
Warning: A single user ArcMap display Desktop casual users every 3-4 years to
session takes advantage of a single maintain work productivity.
processor core. Display performance is
determined by per-core processor

Windows terminal clients and web clients • Quad-core technology is now the
require much less processing, and can work standard for desktop platforms, and
fine with most standard office workstations. although a single ArcMap process
Upgrade terminal and browser client users will see little performance gain in a
every 3-5 years to maintain work multi-core environment there will be
productivity. significant user productivity gains by
enabling concurrent background
Workstation operating system processing loads.
• Turbo boost increases per-core
Full release and support for the Windows processing performance when
64-bit operating system provides supporting a single user session and
performance enhancement opportunities for background processing is not
ArcGIS Desktop workstation environments. required.
Windows 64-bit OS improves ArcGIS • Desktop parallel processing
Desktop memory access, supports a larger environments are leveraged when
number of concurrent background sessions, using a basemap layer or
and takes advantage of higher memory accelerated imagery layer in ArcGIS
capacity for data caching. Working with 10+ applications.
imagery, cached basemap tiles, and local • 3D image streaming with ArcGIS
feature cache are examples where ArcMap Explorer and future enhancements
display performance can be improved when with 3D simulation in ArcGIS Pro
accessing data from local memory cache. and geoprocessing also leverage the
increased capacity of multi-core
Make sure your workstation has sufficient workstation environments.
physical memory to handle your application
workflow. 16 GB memory is adequate for Video display processing
most enterprise geodatabase workflow
clients. 32 GB memory or more may be Video graphics cards enhance the ArcGIS
required when working with large data files Desktop user display environment,
(imagery, shape files, etc). Additional particularly for 3D Analysis performance
memory may be required when working with and imagery display quality, and will be
several different applications on the same particularly important in supporting the new
workstation or working in a virtual desktop ArcGIS Pro desktop application. ArcGIS 3D
environment. Analyst requires OpenGL-compatible
graphics cards, as OpenGL technology is
Workstation performance used for 3D display in the ArcGlobe, City
Engine, and ArcScene applications. ArcGIS
Most GIS users are quite comfortable with Explorer for Desktop also uses OpenGL
the performance provided by current technology for 3D rendering. Frequently
Windows desktop technology. Power users asked questions for selecting a video card is
and heavier GIS user workflows will notice provided in the 3D Analysis ArcGIS Desktop
performance improvements with the faster help documentation.
processor technology.

Windows Terminal Server/Remote Desktop Services Platform Sizing
Windows Terminal Server software

configurations. The ArcGIS Desktop
ArcMap direct connect architecture will be
used to demonstrate how Windows
Terminal Server sizing has been influence
by hardware technology change.
Esri certifies each ArcGIS Desktop release

with Citrix XenApp server (Citrix Receiver)
environment. A more complete discussion
Figure 7.14 Windows Terminal Server Architecture on Centralized Windows Terminal
Windows Terminal Server supports Server/Remote Desktop Services (Citrix)
centralized deployment of ArcGIS Desktop Architecture is provided in Chapter 9.
applications for use by remote terminal
clients. Figure 7.14 identifies three standard
ArcGIS Desktop terminal server platform capacity changes
Figure 7.15 Windows Terminal Server Platform Capacity is Changing
Figure 7.15 identifies how vendor hardware Windows Terminal Server sizing over the
improvements have made a difference in past 6 years. The improvements in

processor core performance in conjunction Terminal Server physical memory
with more processor core per chip provide guidelines
increased server throughput capacity
(number of concurrent users supported on a Servers must be configured with sufficient
single platform). As the number of physical memory to take advantage of the
concurrent user sessions on a platform higher platform processing capacity.
increase, the memory and input/output Heavier user workflows can require more
(storage access) requirements must also session memory than lighter workflows.
increase to accommodate the additional
concurrent user sessions. Heavier It is important to take advantage of the
workflows can require more memory per Windows 64-bit Operating System for the
session than lighter workflows. Servers new Intel platforms, since these higher
must be configured with sufficient physical capacity servers require much more
memory to take advantage of the higher physical memory to handle the high number
platform processing capacity. of current active client sessions. 64-bit
operating systems improve memory
Server 4-core capacity has increased more management and provide up to 10 percent
than 55 percent over the last 6 years. performance gain over Windows 32-bit
Display complexity is increasing for most Server Advanced operating systems.
business workflows, increasing compute
requirements. Following range of users These high performance servers push
support medium to heavy complexity capacity to new levels, and GIS applications
ArcMap remote session workflow patterns. may push platform and disk subsystems to
their limits. Monitor disk traffic and platform
• 2011 4-core server supported 19 - paging during peak loads to ensure these
29 concurrent users (38 GB RAM) subsystems are not overloaded. More
• 2013 4-core server supported 23 - memory can reduce paging and reduce disk
34 concurrent users (46 GB RAM) contention by improving data caching. Data
• 2015 4-core server supported 28 - can be distributed over multiple disk
42 concurrent users (56 GB RAM) volumes to reduce file access contention.
• 2016-2017 4-core server supported You need to know you have a problem
31 - 46 concurrent users (62 GB before you can fix it, so keep an eye on
RAM) platform performance metrics to see all is
• 2016-2017 2-core virtual server working as it should.
supports 13 - 20 concurrent users
(27 GB RAM). CPT for Windows Terminal Server platform
sizing
These are approximate capacity estimates
based on our platform capacity planning CPT can be used for platform sizing. The
models. Recommend using the CPT along recommended platform solution is
with workflow performance guidelines generated by Excel once you enter your
provided in Chapter 3 for proper capacity business requirements and make your
planning. Workflow complexity and use hardware selections. You can try different
productivity can have a significant impact on platform configurations and experiment with
overall platform sizing requirements. different workflow complexities.
The CPT Calculator tab can be used for

single workflow platform sizing. The CPT
Design tab should be used for more detailed
enterprise design planning.

Additional ArcGIS platform memory SDSwiki appendix on Windows Memory
configuration guidelines are provided in the Management.
Enterprise Geodatabase Server Sizing
Figure 7.16 identifies software configuration options for the geodatabase server platforms. The
geodatabase transaction models apply to both ArcGIS Desktop and Web mapping service
transactions. Normally a geodatabase is deployed on a single database server node, and larger
capacity servers are required to support scale-up user requirements.
been influence by hardware technology

change.
Figure 7.17 identifies the impact of

hardware technology change on Enterprise
Geodatabase server sizing over the past 5
years. Improvements in processor core
performance in conjunction with more
processor core per chip have significantly
increased server throughput capacity
(number of concurrent users supported on a
single platform). As the number of
concurrent user sessions on a platform
increase, the memory requirements will
Figure 7.16 Enterprise Geodatabase Server increase to accommodate the additional
Architecture Alternatives concurrent user sessions. Heavier
workflows can require more memory per
The ArcGIS Desktop direct connect session than lighter workflows. Servers
architecture will be used to demonstrate must be configured with sufficient physical
how Enterprise Geodatabase sizing has memory to take advantage of the higher
platform processing capacity.

Enterprise Geodatabase server platform capacity changes
Figure 7.17 Geodatabase Server Platform Capacity is Changing
Database platform capacity has increased productivity can have a significant impact on
by about 66 percent over the past 5 years. overall platform sizing requirements.
• 2011 4-core server supported 191 - Enterprise Geodatabase Platform

287 concurrent users (38 GB RAM) Memory guidelines
• 2013 4-core server supported 230 -
344 concurrent users (46 GB RAM) Servers must be configured with sufficient
• 2015 4-core server supported 281 - physical memory to take advantage of the
422 concurrent users (56 GB RAM) higher platform processing capacity.
• 2016-2017 4-core server supported Heavier user workflows can require more
309 - 463 concurrent users (62 GB session memory than lighter workflows.
RAM)
• 2016-2017 2-core virtual server It is important to take advantage of the
supports 133 - 199 concurrent users Windows 64-bit Operating System for the
(27 GB RAM). new Intel platforms, since these higher
capacity servers require much more
These are approximate capacity estimates physical memory to handle the high number
based on our platform capacity planning of current active client sessions. 64-bit
models. Recommend using the CPT along operating systems improve memory
with workflow performance guidelines management and provide up to 10 percent
provided in Chapter 3 for proper capacity performance gain over Windows 32-bit
planning. Workflow complexity and use Server Advanced operating systems.

These high-performance servers push Identify your peak user requirements and
capacity to new levels, and GIS applications the DB_DBMS data source. Select a two
may push platform and disk subsystems to tier platform architecture and your EGDB
their limits. Monitor disk traffic and platform hardware platform choice.
paging during peak loads to ensure these
subsystems are not overloaded. More The recommended platform solution is
memory can reduce paging and reduce disk generated by Excel once you enter your
contention by improving data caching. Data business requirements and make your
can be distributed over multiple disk hardware selections. You can try different
volumes to reduce file access contention. platform configurations and experiment with
different workflow complexities.
You need to know you have a problem
before you can fix it, so keep an eye on The CPT Calculator tab can be used for
platform performance metrics to ensure all single workflow platform sizing. The CPT
is working as it should. Design tab should be used for more detailed
enterprise design planning.
CPT for ArcSDE Geodatabase platform
sizing Additional ArcGIS platform memory
configuration guidelines are provided in the
CPT Calculator tab can be used for SDSwiki appendix on Windows Memory
Enterprise Geodatabase (EGDB) platform Management.
sizing.
Web Mapping Servers
ArcGIS Server system requirements are provided in the ArcGIS resource center. The legacy
ArcIMS image service was deployed using the ArcIMS software, and the ArcGIS Server map
services are deployed using the ArcGIS Server software (ArcGIS Server is the primary selection
for current Web environments). All Web mapping technologies can be deployed in a mixed
software environment (they can be deployed on the same server platform together). All mapping
services can be configured to access a file data source or a separate Enterprise geodatabase.
Web mapping services have experienced

dramatic performance changes over the
past 5 years. These performance
enhancements improve Web user
productivity and reduce deployment cost.
Some of these performance changes were
due to expanding software deployment
options and others were due to improved
hardware processing speed and platform
capacity changes. Figure 7.18 Web Server Two Tier Architecture
Figure 7.18 identifies recommended

software configuration options for standard
two-tier Web mapping deployments. This
configuration option supports the Web
server and GIS server components on the

same platform tier. Results for four standard Deployment of ArcGIS Server in virtual
two-tier web server software configurations server environments is discussed in the
will be shown. Virtualization deployment options section in
Chapter 9. The Virtual server section shares
• Legacy ArcIMS image service direct some joint testing reports with VMware
connect to an Enterprise Geodatabase. virtualization technology showing ArcGIS
• ArcGIS 10.0 for Server direct connect to Server performance and scalability in
an Enterprise Geodatabase. This was different virtual server deployment
the last release publishing services configurations.
using the windows (MXD) map
rendering engine. Best practice: Selecting the right host
• ArcGIS Server direct connect to an platform for your virtual server
Enterprise Geodatabase. deployment makes a difference -
• Portal for ArcGIS and ArcGIS Server physical host processor performance
direct connect to an Enterprise directly impacts virtual server
Geodatabase and Data Store. performance and scalability
Many organization are deploying ArcGIS

Server in a virtual server environment.
Web mapping server platform capacity changes.
Vendor hardware improvements along with ArcGIS Server architecture changes have made a
difference in web server sizing over the past 5 years.
Figure 7.19 Web mapping server capacity performance changes over the last 5 years

mapping services deployed on a 2-core
Figure 7.19 shows performance changes for virtual server supported up to 31,700
Web services over the past 5 years. transactions per hour. 2017 ArcGIS Server
Medium complexity Web services are platform memory recommendations were
becoming more common. Intel Xeon E5- increased to 4 GB per core, with 8 GB
2643 4-core server was the baseline memory recommended for a 2-core virtual
platform for 2012. ArcIMS image service server. Additional 1 GB per host platform
supported up to 49,500 transactions per core was added for visualization overhead.
hour, ArcGIS Server REST MXD mapping
services supported up to 31,400
ArcGIS Server platform physical memory
transactions per hour, and ArcGIS Server
guidelines
REST MSD mapping services supported up
to 51,400 transactions per hour. Platform
memory recommendations were 3 GB per Servers must be configured with sufficient
core, with 12 GB memory recommended for physical memory to take advantage of the
a 4-core server. ArcGIS Server with new higher platform processing capacity.
MSD map rendering engine outperformed Heavier user workflows can require more
ArcIMS and provided better map quality on session memory than lighter workflows. It is
the same platform environment. important to take advantage of the Windows
64-bit Operating System for the new Intel
platforms, since these higher capacity
Intel Xeon E5-2637v2 4-core server was the
servers require much more physical
baseline platform for 2014. ArcGIS Server
memory to handle the high number of
REST MSD mapping services supported up
current active client sessions. 64-bit
to 60,300 transactions per hour. ArcGIS
operating systems improve memory
Server REST 2D VP mapping services
management and provide up to 10 percent
registered with Portal for ArcGIS supported
performance gain over Windows 32-bit
up to 57,300 transactions per hour. Platform
Server Advanced operating systems.
memory recommendations were 3 GB per
ArcGIS Server deployments should carefully
core, with 12 GB memory recommended for
follow recommended maximum service
a 4-core server.
instance and host platform configuration
guidelines discussed in Chapter 4.
Intel Xeon E3-1270v5 4-core server was the
baseline platform for 2016-2017. ArcGIS
These high performance servers push
Server REST 2D Vector mapping services
capacity to new levels, and GIS applications
supported up to 73,700 transactions per
may push platform and disk subsystems to
hour. ArcGIS Server REST 2D VP mapping
their limits. Monitor disk traffic and platform
services registered with Portal for ArcGIS
paging during peak loads to ensure these
supported up to 70,100 transactions per
subsystems are not overloaded. More
hour. The ArcGIS Server functionality
memory can reduce paging and reduce disk
makes it easy to publish and share map
contention by improving data caching. Data
services and geoprocessing models to a
can be distributed over multiple disk
local server and to ArcGIS Online cloud
volumes to reduce file access contention.
services. 2017 ArcGIS Server platform
memory recommendations were increased
to 4 GB per core, with 16 GB memory You need to know you have a problem
recommended for a 4-core server. before you can fix it, so keep an eye on
platform performance metrics to see all is
working as it should.
Intel Xeon E5-2667v4 16-core server was a
popular virtual server host platform for
2016-2017. ArcGIS Server REST 2D Vector

CPT ArcGIS Server platform sizing
CPT Calculator tab can be used for ArcGIS hardware selections. You can try different
Server platform sizing. platform configurations and experiment with
different workflow complexities.
• Select the workflow description that
represents your user performance The CPT Calculator tab is used for single
targets. workflow platform sizing. The CPT Design
• Identify your peak user requirements tab should be used for more detailed
and your selected data source. enterprise design planning.
• Select your platform architecture and
your hardware platform choice. Additional ArcGIS platform memory
configuration guidelines are provided in the
The recommended platform solution is SDSwiki appendix on Windows Memory
generated by Excel once you enter your Management.
business requirements and make your
Platform Selection Criteria

Figure 7.20 provides a summary of the factors contributing to proper hardware selection. These
factors include the following:
Figure 7.20 Platform Vendor Selection
Esri system design role Platform processor selection: Platform

must be configured properly to support your
User requirements analysis: Proper user performance requirements. Proper
platform selection is driven by your business platform technology selection based on user
needs. You need to know what you need to performance needs and peak system
do before you can identify what you need to processing loads significantly reduces
do it. Once you identify what you need to implementation risk. Esri performance sizing
do, you can use the CPT to identify your models establish a solid foundation for
platform needs. proper hardware platform selection. The
Capacity Planning Tool automates the

System Architecture Design analysis, System Supportability: Customers must
providing a framework for coupling evaluate system supportability based on
enterprise GIS user requirements analysis vendor claims and previous experience with
with system architecture design and proper supporting vendor technology.
platform technology selection.
Vendor Relationships: Relationships with
Network infrastructure requirements: the hardware vendor may be an important
Network bandwidth must be adequate to consideration when supporting complex
support peak throughput loads required to system deployments.
support remote user productivity. Proper
bandwidth is essential to enable Total Life Cycle Costs: Total cost of the
implementation success. The Capacity system may depend on many factors
Planning Tool automates the Network including existing customer administration of
Suitability Analysis to ensure remote site similar hardware environments, hardware
bandwidth requirements are identified for reliability, and maintainability. Customers
projected peak processing loads. must assess these factors based on
previous experience with the vendor
Hardware vendor role technology and evaluation of vendor total
cost of ownership claims.
Purchase Price: Cost of the hardware will
vary depending on the vendor selection and Establishing specific hardware technology
platform configuration. Capacity Planning specifications for evaluation during
Tools can identify specific technology hardware source selection significantly
required to satisfy peak system processing improves the quality of the hardware
needs. Pricing should be based on the selection process. Proper system
evaluation of hardware platforms with equal architecture design and hardware selection
display performance platform workflow provide a basis for successful system
capacity. deployment.

Information Security
Fall 2017 Information Security 41st Edition
This chapter introduces the purpose and scope of information security. Basic concepts are
introduced for developing security solutions that meet your business needs. Esri's information
patterns share how to establish security measures appropriate for your organization.
Enterprise security can be a challenge for IT What is information

architects and security specialists. Until the
last few years, entire IT systems were security?
frequently designed around a single mission
objective and a single "community of
interest," normally supported with physically
isolated systems, each with its own data
stores and applications. New emerging
standards are supported with more mature
communication environments, more
intelligent operating systems, and a variety
of standard integration protocols enabling IT
architects to design and maintain
comprehensive organization-wide
interactive enterprise solutions.
Recent industry advancements, especially

Figure 8.1 Information security is an important part of
in the areas of web service standards and risk management, supports business continuance,
service-oriented architectures, are enabling and is managed in part by information technology.
architects to more effectively satisfy
enterprise security objectives. Esri's careful Information security is the process of
attention to these standards, coupled with protecting the availability, privacy, and
an overall philosophy of providing highly integrity of data. Risk management is an
interoperable software, provides security overall goal of every organization.
architects with a high level of flexibility, thus Information security is one of the disciplines
establishing trust for all Esri components within the organization that addresses risk
contained in an enterprise solution. management. Risk is also managed through
additional business continuance and
A full discussion on enterprise security is information technology initiatives.
beyond the scope of this chapter. The [Trust
ArcGIS] site provides unified access to Information security has some common
security related information for enterprise characteristics with business continuance
solutions using Esri products. The Trust and information technology as shown in
ArcGIS web site includes a section on Figure 8.1.
[ArcGIS Security] for Cloud, Server,
Desktop and Mobile deployment patterns. • Information security is a subset of
overall risk management.
• Information security is important in
maintaining business continuance.
• Information security is managed in
part by information technology.
Information Security 8.1

Four types of security threats
Figure 8.2 Security threats include natural disasters, malicious attacks, internal attacks, and system malfunctions or
human error.
Information security is focused on Best Practice: Security controls are

addressing the four types of security threats developed and deployed to protect
identified in Figure 8.2. These security against identified security threats.
threats include natural disasters, malicious
attacks, internal attacks, and system CIA security triad
malfunctions or human error.
National Institute of Standards and

Technology (NIST) definition of a security
threat: Any circumstance or event with the
potential to adversely impact organizational
operations (including mission, functions,
image, or reputation), organizational assets,
or individuals through an information system
via unauthorized access, destruction,
disclosure, modification of information, Figure 8.3 The CIA security triad provides overall
and/or denial of service. Also, the potential guidance for enterprise security management.
for a threat-source to successfully exploit a
particular information system vulnerability.

Figure 8.3 shows the CIA triad. The core Levels of Security
principals of information security
management are represented by the CIA
triad.
The CIA triad includes confidentiality,

integrity, and availability.
• Confidentiality is protection of
"privileged" communications, restricting
user access to core business
information based on a "need to know"
principle.
• Integrity refers to the trustworthiness of Figure 8.4 Several levels of security are required to
business data resources and the ensure protected business operations.
associated information products
generated over its entire life cycle. Figure 8.4 shows the defense in depth
• Availability refers to ensuring the concept. Defense in depth is an information
information system is functional when assurance concept in which multiple layers
needed to support operational business of security controls (defenses) are placed
requirements. throughout an IT system.
Information security industry standards will Multiple levels of security:

be identified and applied as mechanisms of
protection and prevention in the following • Physical controls (fences, guards, locks,
three main areas: etc.)
• Policy controls (administrative policies
• Hardware and procedures)
• Software • Technical controls (system
• Communications configuration)
Protection and prevention will be Types of technical controls:

implemented at three levels, or layers:
• Authentication (user identity strategy,
• People (personal security) user name and password, keycards,
• Procedures (organizational security) keywords, etc.)
• Products (physical security) • Authorization (role-based access
policies, access control rules, etc.)
Best practice: The CIA triad is used to • Filters (routing based on group policy,
provide proper scope and focus for active directory containers, user identity,
information security management. etc.)
• Encryption (scrabbling information for
unreadable transmission or storage)
• Logging (record of security-related
transactions)

Technical controls are implemented Defense in depth seeks to delay advance of
throughout the physical system providing an attack:
multiple layers of defense:
• Yield space in order to buy time without
• Application controls (LDAP, SSO, HTML preventing proper access.
content filters, validation checks, secure • Prevent penetration and direct attacks
stored procedures) by providing multiple layers of defense.
• Host/device controls (native • Prevent security breaches and buy time
authentication, LDAP, repository, to detect and respond to an attack.
hardening guides, HIDS)
• Network controls (firewalls, NIDS, single Best practice: Multiple layers of defense
socket layer - SSL, IPSEC) improve information security.
• Data controls (authentication, role-
Warning: Do not expect a high level of
based authorization, row-level access,
protection from a single layer of defense.
data file encryption)
Examples of defense in depth: Enterprise security strategy
• Application functional limitations (view Review current security trends

only)
• Reverse proxy server (restrict port Information security is a growing science
access)
• Web application firewall (monitor traffic, • Dollar amount losses by threat
restrict access, route traffic) • Security technologies utilized
• Web server (provide extra physical
transmission layer) Review security options
• ArcGIS Server (restrict access to
published services, user authentication, • Trust ArcGIS site
restricted data access) • Enterprise-wide security mechanisms
• Geodatabase server (restrict access to • Application-specific options
published services, user authentication,
restricted table and row access, monitor Implement security as a business enabler
traffic)
• Improve appropriate availability of
Idea behind defense in depth: information.
• Defend a system using multiple varying

protection methods.
• Provide a comprehensive approach to
information security.

Standards approach to security risk management.
Figure 8.5 Security risk management process diagram.
Figure 8.5 shows a standards approach to measurable, attainable, relevant, and

security risk management. Standard time-bound.
approaches to security risk management 4. Framework. Identify an information
are well established and should be followed security management approach and
to ensure compliance. methodology that will deliver results.
Several frameworks have been
Identify your security needs developed and shared for general use in
establishing an information security
• Review industry security threats. program. Information Security
• Assess your environment. Frameworks can be industry specific
• Evaluate risk to datasets and and share focused best practices that
operational systems. address your business needs.
• Determine sensitivity, categorization, 5. Approved planning. Establish a plan for
and patterns of risk. the security risk assessment effort. You
will need management authorization for
Key steps to effective information security: required resources, support, and
funding.
1. Legislation. Review regulations related 6. Risk assessment and mitigation.
to your industry. Security regulations Complete a risk assessment security
may dictate compliance standards and a needs analysis identifying potential
security implementation framework; threats and associated mitigation
there may be negative business strategies.
consequences for non-compliance. 7. Safeguards. Identify security procedures
2. Benefits. Identify any potential benefits (rules) and technology (tools) that must
that can be derived from security be implemented to address identified
compliance and operational savings that security needs.
can be attributed to the proposed 8. Training and awareness. Design and
security program. This can be helpful in build the approved security solutions.
justifying security program expenses. Implement training and awareness
3. Objectives. Establish SMART programs to implement and enforce
information security program objectives. identified security practices.
Objectives should be specific,

9. Implementation. Operate and support IT group must be on board to design, build,
the security solutions. Monitor levels of and operate the approved security
protection and measure compliance. solutions. Periodic audit reviews and formal
compliance demonstrations are essential to
Steps 1-5 are required efforts to establish a assess risk management effectiveness.
successful security program. Security Executive sponsor should actively review
program should be presented and endorsed progress in meeting the established security
by Executive management, and an program SMART objectives.
Executive sponsor should be actively
committed to enforcing the objectives of the Best practice: Security management is a
security program. continuous process of reviewing and
updating security rules and supporting
Information security management is an technology to maintain a proper level of
active ongoing effort assessing risks, defense against evolving security
defining security requirements, and threats.
measuring security solutions.
Esri informal pattern selection
Figure 8.6 Classify your security needs based on your own security risk.
Your security needs are unique. Figure 8.6 shows a full range of security levels available for
ArcGIS users. Esri provides an approach to classifying the level of security required to manage
your security risk.

Basic security Esri security strategy
• Minimum level of security investment. evolution
• Enables simple and lowest system cost.
• Enables full access to internet data
sources and Online services.
• Provides optimum business
environment for external collaboration.
• Extends enterprise operations to include
connected mobile applications.
• Protects system from internet virus
attacks.
Standard security
• Moderate level of security investment.

• Moderate increase in complexity and Figure 8.7 GIS security solutions are driven by
system cost. technology change, following the patterns common to
• Enables full access to Internet data most standard IT trends.
sources and online services.
• Provides optimum business Figure 8.7 shows security moving from an
environment for external collaboration. isolated product solution focus to
• Extends enterprise operations to include addressing security on an integrated
connected mobile applications. solutions level. Enterprise IT solutions are
• Protects system from a variety of changing including more transparency,
security risks. sharing, collaboration, and web access.
Security policies are adapting to these
Advanced security changes.
• Heavy level of security investment. Product

• High increase in complexity and system
cost. User workflow environment:
• Restricts access to Internet data
sources and online services. • Isolated systems
• Eliminates external online collaboration. • Primarily desktop or internal network
• Prevents most connected mobile solutions
applications. • Limited web access
• Provides optimum protection to manage • Data entry provided by well-defined
security risks. workflows
Best practice: Apply appropriate Security solutions focused on isolated

mitigation strategies to address your systems:
unique confidentiality, integrity, and
availability business requirements. • Protecting discrete products and
services
• Protecting focused user workflow
environments
• Include third-party security additions

Enterprise • Adaptive user-driven security controls
• Include third-party security additions
User workflow environment:
Solution
• Multiple clients and user locations
• Multiple servers and data center Managed security solutions:
locations
• User collaboration across multiple • Solution templates established based on
integrated systems industry standards.
• Discretionary user grouping and sharing • Best practices developed and shared by
• Common interface with cloud-hosted community leaders.
services • Solution strategies involve integration of
multiple enterprise environments.
Enterprise security solutions: • Security solutions are expanded to
include cloud deployments.
• Integrated enterprise platforms and
services
• Multi-layered embedded security
protection
ArcGIS Server: Authorization deployment scenarios
Figure 8.8 ArcGIS Server provides product and enterprise level security options.
Figure 8.8 shows the security options

available with ArcGIS Server. ArcGIS ArcGIS Server supports both Server tier and
Server provides three primary options for Web tier integration with Active Directory or
managing user identify and access control. LDAP directory store for enterprise-level
management of user identity and service
Product-level security management access control. User access to secured
services can be managed by the Enterprise
ArcGIS Server includes a built-in identity security administrator.
store for local management of user identity
and service access control. User access to Best practice: Business requirements
secured services can be managed by the will determine the optimum security
ArcGIS Server site administrator. management solution for your
organization.
Enterprise-level security management

ArcGIS Server tier authentication
Figure 8.9 ArcGIS Server tier authentication.
Figure 8.9 shows the ArcGIS Server built in o Service publishers share
security option. published services with available
roles.
ArcGIS Server (AGS) site manages user • Service authorization is provided by
authentication and service authorization. ArcGIS Server token based
authentication.
• Authentication credentials are stored in
the AGS site identity store (secure ArcGIS Server tier security authorization
users). data flow
• Authorization credentials are stored in
the AGS site identity store (roles). • User security credentials are provided to
• Privileges are assigned by the AGS site Web adaptor or third party
administrative user roles. reverse proxy.
• Service access is authorized based on • AGS site Web adaptor (or third party
identified roles (AGS folders). reverse proxy) sends credentials to the
o ArcGIS Server administrator ArcGIS Server site.
manages user membership, user • AGS site identity store is used to
privileges and service access complete authentication and
permissions (roles), and assigns authorization
users to roles.

• Service authorization is provided by Help: Configuring ArcGIS Server security
ArcGIS Server token based
authentication.
Enterprise ArcGIS Server tier authentication
Figure 8.10 ArcGIS Server Enterprise level Web tier authentication.
Figure 8.10 shows the enterprise level o ArcGIS Server administrator

ArcGIS Server tier security option. manages user privileges and
access permissions (roles), and
ArcGIS Server (AGS) site manages service assigns Active Directory/LDAP
authorization based on validated enterprise identified users to roles.
user authentication. o Service publishers share
published services with available
• Authentication credentials are stored in roles assigned by the ArcGIS
the enterprise Active Directory/LDAP Server site administrator.
data store (secured users) and
replicated to the ArcGIS Server identity • Privileges managed by Active
store (user name and password). Directory/LDAP data store (roles).
o Enterprise Active Directory/LDAP
Two authentication options are available. administrator manages user
membership, user privileges and
• Privileges managed by ArcGIS Server access permissions (roles), and
site identity store (roles). assigns users to roles.
o Enterprise Active Directory/LDAP o Service publishers share
administrator manages user published services with available
membership.

roles assigned by the Active • AGS site identity store is used to
Directory/LDAP administrator. complete authentication and
• Service authorization is provided by authorization
ArcGIS Server token based • Service authorization is provided by
authentication. ArcGIS Server token based
authentication.
Enterprise ArcGIS Server tier security
authorization data flow Help: Advanced considerations when using
domain accounts
• ArcGIS Server site identity store read
only trust relationship is configured with Best practice: Use secure socket layer
the Enterprise security directory data (SSL) communications when
store. transmitting user identification
• User security credentials are provided to information over unsecure network.
the AGS site Web adaptor or third party
reverse proxy.
• AGS site Web adaptor (or third party
reverse proxy) sends credentials to the
ArcGIS Server site.

Enterprise level Web tier authentication
Figure 8.11 ArcGIS Server Enterprise level Web tier authentication.
Figure 8.11 shows the ArcGIS Server Web • Privileges managed by Active
tier security option. ArcGIS Server (AGS) Directory/LDAP data store (roles).
site manages service authorization based o Enterprise Active Directory/LDAP
on validated enterprise user authentication. administrator manages user
membership, user privileges and
• Authentication credentials are stored in access permissions (roles), and
the enterprise Active Directory/LDAP assigns users to roles.
data store (secured users). o Service publishers share
published services with available
Web-tier authentication supports user single roles assigned by the Active
sign-on experience. Directory/LDAP administrator.
• Service authorization is provided by
Two authentication options are available. ArcGIS Server token based
authentication.
• Privileges managed by ArcGIS Server
site identity store (roles). Enterprise level Web tier security
o Enterprise Active Directory/LDAP authorization data flow
administrator manages user
membership. • ArcGIS Server site identity store
o ArcGIS Server administrator read only trust relationship is
manages user privileges and configured with the Enterprise
access permissions (roles), and security directory data store.
assigns Active Directory/LDAP • User security credentials are
identified users to roles. provided to the Web server.
o Service publishers share • Web server sends user credentials
published services with available to the Active Directory/LDAP server.
roles assigned by the ArcGIS • Active Directory/LDAP data store is
Server site administrator. used to complete authentication.

• Validated authentication credentials Help: Securing web services with Integrated
are returned to the Web server. Windows Authentication
• AGS Web Adaptor sends validated
authentication credentials to the Best practice: Use secure socket layer
ArcGIS Server site. (SSL) communications when
• GIS Server identity store provides transmitting user identification
authorization for service access to information over unsecure network.
client.
ArcGIS Server token based
authentication.
Portal: Authentication deployment scenarios
Figure 8.12 ArcGIS Online and Portal for ArcGIS support product-level and enterprise-level security authentication.
Figure 8.12 shows the security options Enterprise-level security management

available for ArcGIS Online and Portal for
ArcGIS deployments. Both portal • ArcGIS Online and Portal for ArcGIS
deployment patterns support product-level both support SAML integration with
and enterprise-level user identity Active Directory or LDAP directory
management. store.
• Portal for ArcGIS supports both
Product-level security management Portal tier and Web tier integration
with Active Directory or LDAP
• ArcGIS Online includes a built-in directory store.
global security store.
• Portal for ArcGIS includes a built-in Portal content is created and shared by
identity store. named users using the portal information
model. Security is managed by the ArcGIS
Portal content is created and shared by Online or Portal for ArcGIS administrator.
named users using the portal information User access to secured services can be
model. Security is managed by the ArcGIS managed by the Enterprise security
Online or Portal for ArcGIS administrator. administrator.

Best practice: Business requirements The GIS portal gives a self-service content
will determine the optimum security management platform for managing
management solution for your geospatial content as shown in Figure 8.13.
organization.
The portal information model includes
ArcGIS portal information model Users, Groups, Items, and Tags.
• Users own items and can own or join

groups.
• Groups are used to organize and secure
user items.
• Items identify user content added to the
Portal.
• Tags identify item content for search
purposes.
Figure 8.13 The portal information model connects

users with geospatial content for collaboration and
sharing.
Web GIS access and privileges
Figure 8.14 Portal privileges available with ArcGIS Online for Organizations.
Portal privileges are based on named user Administrator has the capability to restrict
roles managed by the Portal administrator shared services to named users within the
as shown in Figure 8.14. Published maps organization (exclude anonymous access).
and apps can be shared to anonymous
(public) users outside the organization.

Portal security is managed by named user and Community demographics,
membership with the following privileges: spatial analysis, network routing,
world geocoding, and landscape
• Administrators have full permissions feature services. Use of these
and manage portal named user services consume Organization
membership. online credits.
• Named users with Viewer role can • Administrators are also able to
use maps and apps and join groups. create custom roles for more
• Named users with Contributor role focused named user privileges.
can also create content, share maps ArcGIS Online custom roles allow
and apps, join and create groups, administrators to assign granular
and edit features. access (specific online analysis
• Named users with a publisher role services) to named users based on
can also publish hosted web layers their operational needs.
and have full access to ArcGIS
Online analysis services. Online
services include access to Business
ArcGIS Online security authentication and authorization

ArcGIS Online provides secure access to • Organization named users can
shared maps, apps, and data packages create and manage their own group
hosted in your private ArcGIS Online membership and permissions.
Organization in the Cloud. Organization Groups can be private, organization,
membership is limited to named users, with or public access. An Online
member authentication and resource Organization membership is
access managed in a Cloud based security required to participate in managed
store. Security assertion markup language group membership. Group users
(SAML) authentication can be used to have contributor or viewer
integrate the ArcGIS Online Organization permissions.
security store with on-premise security • When you add layers to an ArcGIS
solutions for Enterprise level member Online Web map from a Web
authentication. service, these layers are published
from their source site and delivered
Groups are created and managed by direct to the client. The source site
Organization named users. manages any additional client
authentication and validation
• Administrators have full permissions requirements for the selected
and manage organization service (no data gets transferred
membership (named users) and through the ArcGIS Online site.)
user roles. Web map layers are assembled in
the mashup at the client browser
display.

ArcGIS Online authentication
Figure 8.15 ArcGIS Online Global Security Store authentication.

Figure 8.15 shows ArcGIS Online Global • Service publishers share published
Security Store authentication. services with ArcGIS Online
Organization groups.
ArcGIS Online global security store • Service access authorization is based
manages authentication and authorization. on group membership.
• Authentication credentials are stored in ArcGIS Online security authorization data

the ArcGIS Online global security store flow.
(named users).
• Privileges and group membership are • User security credentials are provided to
stored in the ArcGIS Online global the ArcGIS Online global security store.
security store (roles). • ArcGIS Online global security store is
• ArcGIS Online Organization used to complete authentication and
administrator identifies and manages authorization
named users, creates custom privileges, • Service authorization is provided by
and assigns named user privileges ArcGIS Online token based
(roles). authentication.
• ArcGIS Online named users create and
manage ArcGIS Online Organization
groups.

ArcGIS Online SAML authentication
Figure 8.16 ArcGIS Online Federated SAML authentication.
Figure 8.16 shows ArcGIS Online ArcGIS Online for Organizations SAML
Federated SAML authentication. Active security authorization data flow
directory or LDAP can be used for Online
Organization membership authentication. • ArcGIS Online global security store
SAML communication protocols are used SAML identify provider trust relationship
for remote enterprise-level member is configured with the Enterprise security
authentication and validation. directory data store.
• User security credentials are provided to
ArcGIS Online service authorization based the SAML identify provider.
on SAML authentication. • SAML identify provider sends user
credentials to the Enterprise Active
• Authentication credentials are stored in Directory/LDAP server.
the enterprise Active Directory/LDAP • Active Directory/LDAP data store is
data store (named users). used to complete authentication.
• Enterprise Active Directory/LDAP • Validated authentication credentials are
administrator identifies and manages shared with the ArcGIS Online global
ArcGIS Online Organization named user security store.
membership. • Global security store credentials are
• ArcGIS Online server administrator used to authorize named user privileges
defines custom privileges, assigns and services access.
privileges to identified SAML validated • AGOL global security store provides
users, and manages ArcGIS Online authorization for service access to
groups. client.
• ArcGIS Online named users create and • Service authorization is provided by
manage ArcGIS Online Organization ArcGIS Online token based
groups. authentication.
• ArcGIS Online named users share
published services with identified Best practice: Use secure socket layer
ArcGIS Online Organization groups. (SSL) communications when
• Service access authorization is based transmitting user identification
on group membership. information over unsecure network.

Portal for ArcGIS security authentication and authorization
Portal for ArcGIS security solutions provide level integration using Active
Portal tier authentication by the Portal for Directory/LDAP authentication.
ArcGIS identity store or Web tier Enterprise
Portal for ArcGIS server authentication
Figure 8.17 Portal for ArcGIS server authentication.
Figure 8.17 shows Portal for ArcGIS server • Service access authorization is
authentication. based on group membership.
Portal for ArcGIS identity store manages Portal for ArcGIS server tier security
authentication and authorization. authorization data flow
• Authentication credentials are stored in • User security credentials are provided to

the Portal identity store (named users). the Portal Web adaptor.
• Privileges and group membership are • Portal Web adaptor sends user
stored in the Portal identity store (roles). credentials to the Portal server.
o Portal server administrator • Portal identity store is used to complete
identifies and manages named authentication and authorization
users, creates custom privileges, • Service authorization is provided by
and assigns named user Portal for ArcGIS token based
privileges (roles). authentication.
o Portal named users create and
manage Portal groups. Help: About configuring portal
o Service publishers share authentication
published services with groups.

Enterprise Portal for ArcGIS tier authentication
Figure 8.18 Enterprise Portal for ArcGIS tier authentication.
Figure 8.18 shows Enterprise Portal for assigns privileges to identified

ArcGIS tier authentication. Active Directory/LDAP users,
and manages Portal groups.
Portal for ArcGIS manages service o Portal named users create and
authorization based on validated enterprise manage Portal groups.
user authentication. o Portal named users share
published services with identified
• Authentication credentials are stored in Portal groups.
the enterprise Active Directory/LDAP
data store (named users) and replicated • Portal named users and privileges
to the Portal for ArcGIS identity store. (roles) are identified and managed by
the Enterprise Active Directory/LDAP
Three unique security management administrator.
configurations options. Organization must o Enterprise Active Directory/LDAP
select the security management option that administrator identifies and
best supports their business needs. manages Portal named user
membership, defines custom
• Portal named users are identified and privileges, and assigns privileges
managed by the Enterprise Active to identified Portal named users.
Directory/LDAP administrator. o Portal server administrator
o Enterprise Active Directory/LDAP manages Portal groups.
administrator identifies and o Portal named users create and
manages Portal named user manage Portal groups.
membership. o Named users share published
o Portal server administrator services with identified Portal
defines custom privileges, groups.

• Portal named users, privileges (roles), Portal for ArcGIS Web tier security
and Portal groups are identified and authorization data flow
managed by the Enterprise Active
Directory/LDAP administrator. • Portal for ArcGIS identity store read only
o Enterprise Active Directory/LDAP trust relationship is configured with the
administrator identifies and Enterprise security directory data store.
manages Portal named user • User security credentials are provided to
membership, defines custom the Portal Web adaptor.
privileges, assigns privileges to • Portal Web adaptor sends user
identified Portal named users, credentials to the Portal server.
and manages Portal groups. • Portal identity store is used to complete
o Portal named users share authentication and authorization
published services with identified • Service authorization is provided by
Portal groups. Portal for ArcGIS token based
• Service access authorization is based authentication.
on group membership.
Help: Configure Portal to use Enterprise
Identity Store

Portal for ArcGIS Enterprise level Web tier authentication
Figure 8.19 Portal for ArcGIS Web tier authentication.
Figure 8.19 shows Portal for ArcGIS Web assigns privileges to identified
tier authentication. Active Directory/LDAP users,
and manages Portal groups.
Portal for ArcGIS manages service o Portal named users create and
authorization based on validated enterprise manage Portal groups.
user authentication. o Portal named users share
published services with identified
• Authentication credentials are stored in Portal groups.
the enterprise Active Directory/LDAP • Portal named users and privileges
data store (named users). (roles) are identified and managed by
the Enterprise Active Directory/LDAP
Three unique security management administrator.
configurations options. Organization must o Enterprise Active Directory/LDAP
select the security management option that administrator identifies and
best supports their business needs. manages Portal named user
membership, defines custom
• Portal named users are identified and privileges, and assigns privileges
managed by the Enterprise Active to identified Portal named users.
o Portal server administrator
Directory/LDAP administrator.
o Enterprise Active Directory/LDAP
manages Portal groups.
o Portal named users create and
administrator identifies and
manages Portal named user manage Portal groups.
o Named users share published
membership.
o Portal server administrator
services with identified Portal
defines custom privileges, groups.

• Portal named users, privileges (roles), with the Enterprise security directory
and Portal groups are identified and data store.
managed by the Enterprise Active • User security credentials are
Directory/LDAP administrator. provided to the Web server.
o Enterprise Active Directory/LDAP • Web server sends user credentials
administrator identifies and to the Active Directory/LDAP server.
manages Portal named user • Active Directory/LDAP data store is
membership, defines custom used to complete authentication.
privileges, assigns privileges to • Validated authentication credentials
identified Portal named users, are returned to the Web server.
and manages Portal groups. • Portal Web adaptor send validated
o Portal named users share authentication credentials to the
published services with identified Portal server.
Portal groups. • Portal identity store provides
• Service access authorization is based authorization for service access to
on group membership. client.
Portal for ArcGIS Web tier security Portal for ArcGIS token based
authorization data flow authentication.
• Portal for ArcGIS identity store read Help: Using Integrated Windows
only trust relationship is configured Authentication with your portal

Portal distributed collaboration
Figure 8.19.1 Portal to Portal communications are managed through trusted site configurations.
Figure 8.19.1 shows a Portal to Portal Distributed collaboration options include

collaboration architecture, providing a synchronizing Portal group content and
capability for linking Portal for ArcGIS copy/synchronizing Portal group feature
Groups to a collaboration workspace. layer content.
Portal administrators establish trust Security in the cloud

relationship between Portals and create
groups that share with the collaboration
workspace.
Each Guest Portal administrator can

configure when content is synchronized with
the collaboration workspace.
• Immediately
• Default setting
• When item is first shared with the group
• When the shared item is updated
Default setting can be defined as Figure 8.20 Security in the cloud is one of the
challenges facing security managers as they leverage
appropriate for business requirements. new architecture capabilities.
• Every 24 hours by default Figure 8.20 shows the standard Cloud

• At a scheduled interval (minimum hosting patterns and user security practices.
interval is hourly) Security challenges in the cloud are familiar
• Modified using the Sharing API to any IT manager: loss of data, threats to

the infrastructure, and compliance risk. • A closed-perimeter approach or a "fill-
What is new is the way these threats play the-gap" measure. Organizations can no
out in a cloud environment. longer rely on firewalls as a single point
of control, and cobbling together
ArcGIS in the cloud security solutions to protect a single
vulnerability may leave you open in
Software as a Service (SaaS): Direct user places you do not suspect.
interface for building services • Something you can assume is provided
at the level you require by your cloud
• ArcGIS Online (ArcGIS.com) service providers. Make sure you spell
• Business Analyst Online out and can verify what you require.
• ArcGIS Explorer Online
Warning: Cloud computing security is a
Platform as a Service (PaaS); Developer broad topic with hundreds of
interface for building services considerations: from protecting
hardware and platform technologies in
• Esri web mapping APIs (JavaScript, the data center to enabling regulatory
Flex, Silverlight) compliance and defending cloud access
• Microsoft Azure ArcGIS applications through different end-point devices.
Infrastructure as a Service (IaaS): IT Web GIS deployment scenarios

administrator interface for building services
• ArcGIS on Amazon EC2

• ArcGIS on Microsoft Azure
• Private cloud
Cloud security is:
• The response to a familiar set of

security challenges that manifest
differently in the cloud.
• A set of policies, technologies, and
controls designed to protect data and
infrastructure from attack and enable Figure 8.21 ArcGIS deployment options are
regulatory compliance. expanding to include cloud platforms. Security
• Layered technologies that create a management options vary based on the available
durable security net or grid. service models, deployment models, and
management models utilized in your deployment
• The joint responsibility of your scenario.
organization and its cloud provider(s).
Cloud security is evolving to satisfy
Cloud security is not: customer needs.
• A one-size-fits-all solution that can Several deployment strategies are available

protect all your IT assets. In addition to that can include a mix of self-managed and
different cloud delivery models, the vendor-managed security options.
cloud services you deploy will most
likely require more than one approach to
security.

Self-managed deployment options: Deployment strategies can include a mix of
customer-managed and vendor-managed
• Non-cloud on-premises ArcGIS Server security options.
deployment
• Portal for ArcGIS on-premise content Vendor-managed deployment options:
management
• Vendor-managed SaaS-based services
Vendor-managed deployment options: deployment (ArcGIS Online
Organization)
• Vendor-managed hybrid deployment • Vendor-managed IaaS-based
that includes IaaS public and private community private cloud deployment
cloud services • Vendor-managed hybrid deployment
• Vender-managed ArcGIS Online content including IaaS public and private cloud
management services
• Public SaaS-based services deployment • Vendor-managed IaaS-based public
cloud deployment
Best practice: Platform deployment • Vendor-managed ArcGIS Online content
strategy involves a proper trade-off management
between self- and vendor-managed risk.
Customer-managed deployment options:
• Non-cloud on-premise ArcGIS Server

Web GIS deployment options deployment
• Portal for ArcGIS on-premise content
management
• IaaS-based community private cloud
deployment
• Hybrid deployment including IaaS and
on-premise services
Optimum Web GIS deployment includes a

proper mix of customer-managed and
vendor-managed security policies.
Figure 8.21 Web GIS deployment options are

Best practice: An optimum security
expanding to include Cloud platforms. program involves an appropriate tradeoff
between customer- and vendor-managed
Cloud security is evolving to satisfy risk.
customer deployment needs. Figure 8.21
shows the options available for Web GIS
deployment. Security may be a primary
factor in determining the optimum
deployment strategy. Security management
options vary based on the available service
models, deployment models, and
management models utilized in your
deployment scenario.

Deployment model responsibilities managed and vendor-managed security
options.
Hybrid solutions leverage the best

technology options:
• Internal-hosted service layers can

provide your full internal level of
security.
Figure 8.22 System operations management • Private IaaS clouds provide scalable on-
responsibilities vary based on the selected demand internal services while retaining
deployment pattern. required security.
• Geodatabase replication services
Figure 8.22 shows responsibility by layers provide filtered content to physically
across the major cloud deployment models separate internal secure data from
versus an on-premise implementation. external remote access.
• Sensitive data layers can be published
These deployments are not exclusive, and from within the data center for mash-up
an enterprise deployment of the ArcGIS with authenticated field-worker displays.
platform could use multiple models such as • ArcGIS Online organizations provide
an on-premise implementation collaboration and data sharing with
supplemented with ArcGIS in the cloud in a protected groups of agency locations.
hybrid approach. • ArcGIS Online subscription services
(and public IaaS hosting) provide
Best practice: Security implementation scalable public access for on-demand
involves a proper trade-off between self- services.
and vendor-managed risk.
Best practice: Provide optimum
ArcGIS cloud hybrid capabilities enterprise security through hybrid cloud
deployments.
Esri’s security strategy

Deliver secure GIS products
• [Incorporate security industry best

practices.]
• Trusted geospatial services across the
globe.
Figure 8.23 The optimum GIS deployment will often • Meet needs of individual users and
include a mix of internal on-premise, private IaaS, and entire organizations.
Online hosted services.
Provide secure GIS solution guidance
Figure 8.23 shows a ArcGIS Server hybrid
cloud deployment. A hybrid cloud may • [Enterprise Resource Center]
provide your best deployment solution, • [Shares Esri security patterns]
taking advantage of available technology in
the most optimum way without
compromising security. Deployment
strategies can include a mix of self-

Enterprise security firewall patterns
Web firewall best practices
Figure 8.24 The web firewall configuration practices depend on your security needs and proper technology
deployment.
Figure 8.24 shows best practices for firewall

protection. Firewall configurations are Public services should be deployed on
provided to support communication between separate servers from sensitive private
various levels of security. The effectiveness internal services.
of your firewall configuration will depend on
proper technology implementation. • Separate web services tier increases
security layer protection.
Esri provides guidance and • Deploy public services and internal
recommendations for different security private services on separate GIS server
patterns based on your security needs. sites.
• Separate publication dataset from
• [Enterprise GIS Security Patterns] production dataset for optimum
• [Configuring ArcGIS 10.2 for Server protection.
security]
• [Ports used by ArcGIS 10.2 for Server.] High-availability services avoid a single
• [Ports used by Portal for ArcGIS 10.2.] point of failure.
Best practice: Security in depth provides • Multiple servers ensure operational

multiple layers of defense between system with one server down.
public access and protected data • Multiple online copies of operational
resources. data ensure continued operations with
loss of one copy.

• Point-in-time backups are critical—most • Service instance executables access
data corruptions are caused by required data sources and service the
procedural error. request.
• Additional backup copy of critical data • Service instance output is delivered
should be stored off-site. back to the web client.
Additional functionality
Web services with proxy server
• Application firewall features can protect
against common web-based attacks.
Without a reverse proxy, removing
malware or initiating takedowns, for
example, can become difficult.
• In the case of secure websites, the SSL
encryption is sometimes not performed
by the web server itself, but is instead
off-loaded to a reverse proxy that may
be equipped with SSL acceleration
hardware.
• A reverse proxy can distribute the load
from incoming requests to several
Figure 8.25 A reverse proxy server provides a
minimum level of protection for web publishing. servers, with each server serving its
own application area. In the case of
Figure 8.25 shows ArcGIS web services reverse proxying in the neighborhood of
with proxy server. Reverse proxy servers web servers, the reverse proxy may
hide the existence and characteristics of the have to rewrite the URL in each
internal application server. incoming request, in order to match the
relevant internal location of the
requested resource.
Best practice: Basic security: Internal
• A reverse proxy can reduce load on its
web server components can be installed
origin servers by caching static content,
on a single server tier to reduce cost.
as well as dynamic content. Proxy
caches of this sort can often satisfy a
ArcGIS Server reverse proxy architecture
considerable amount of website
(ArcGIS 10.1+):
requests, greatly reducing the load on
the origin server(s). Another term for this
• Web client sends request to web server is web accelerator.
in the DMZ. • A reverse proxy can optimize content by
• DMZ web server sends request to compressing it in order to speed up
reverse proxy for routing to private GIS loading times.
servers. • Reverse proxies can be used whenever
multiple web servers must be accessible
Best practice: ArcGIS Server web via a single public IP address. The web
adaptor will provide reverse proxy and servers listen on different ports in the
load balancing to the private GIS server same machine, with the same local IP
site. address or, possibly, on different
machines and different local IP
• GIS server distributes (load balances) addresses altogether. The reverse proxy
in-bound requests to available service analyzes each incoming call and
instance located within the GIS server delivers it to the right server within the
site. local area network.

GIS server access to required data sources
Web and ArcGIS Server components must be secured.
in DMZ
• File sources must be replicated to the
DMZ to protect internal resources.
• DBMS data sources should be
replicated to DMZ for optimum security.
Warning: Some security officers find this

solution not acceptable because it
provides direct access to the DBMS from
the DMZ.
• SSL secured port connections can be

used to access internal DBMS data
source.
Figure 8.26 Web and GIS server components
deployed in the DMZ. Best practice: ArcGIS Server web
adaptor will provide reverse proxy and
Figure 8.26 shows ArcGIS web and GIS network load balancing protecting
Server components in the DMZ. Web and administrative access to the GIS server
GIS server components can be deployed in site.
the DMZ along with replicated data sources
or with access through the firewall to an • GIS server distributes (load balance) in-
internal DBMS. bound requests to available service
instance located within the GIS server
Best practice: Basic security: Web site.
• Service instance executables access
server DMZ components can be installed
on a single server tier to reduce cost. required data sources and service the
request.
• Service instance output is delivered
ArcGIS Server reverse proxy architecture
back to the web client.
(ArcGIS 10.1+):
Best practice: Web application server
• Reverse proxy secures administrative
installed with the web server can
access to GIS server.
enhance web service security.
• Web adaptor provides reverse proxy
and network load balancing.
• Web application firewall can enhance
web service security.

Business continuance
Business continuance addresses infrastructure design considerations that ensure computer
systems are functional when needed to support operational business requirements.
Requirements for business continuance are a primary risk management consideration, and a
core principal of information security management. Availability is the third tenant of the
Security Triad, and system component redundancy is the principal concept developed and
deployed to protect against identified business continuance threats.
The level of protection required by any Requirements for high-availability

particular business operation is a balance deployments
between cost of the supporting systems and
the risk of downtime. Basic levels of
protection can be satisfied by incremental
data backups and off-site storage. Standard
levels of protection would include high
availability solutions avoiding downtime
caused by a single system component
failure. Advanced levels would include
disaster recovery to a secondary data
center with the loss of the primary site.
Recovery time from a protected failure
event will depend on the capabilities of the
high available and/or disaster recovery
solution. There is no absolute answer that Figure 8.27 High-availability configurations support
continued business operations during single
establishes the level of protection required component failures.
to manage the risk of system downtime.
Figure 8.27 shows required high-availability
ArcGIS Server and Web GIS (component configuration support for continued business
and system level) business continuance operations. High-availability configurations
deployment strategies will be addressed in ensure business operations continue when
this section. Business continuance infrastructure components fail.
deployment options will address product
level high-availability and disaster recovery General high-availability compliance
deployment scenarios. Requirements are requirements include the following:
established based on business security risk
management objectives. Software
• Highly available load balancer
functionality can introduce constraints on
• Multiple application server tier machines
acceptable deployment options.
• Highly available shared storage
• Highly available network
Design and deployment of the required
deployment architecture patterns are
For virtual server environments, high
addressed in the earlier chapter on GIS
availability platform solutions (clustered or
Product Architecture.
failover virtual machines) must be deployed

on separate host platforms to avoid single • Multiple site service deployments
point of failure. • Multiple site data source updates
• Concurrent application updates between
Best practice: System designed to sites
support continued operations following
any single hardware component failure. Deploying effective disaster recovery
solutions can be very complicated and
require enterprise level IT management and
Disaster recovery: Typical workflow governance. The most effective solutions
are supported by virtualization software
technology, with VMware as one of the
leading software vendors in the
virtualization marketplace. Virtualization
software can support a variety of business
systems through a common set of
architecture solutions, managed by
enterprise level IT operations with proper
training and staff to ensure compliance with
business requirements. The VMware
vCenter Server Availability Guide shares
Figure 8.28 Disaster recovery solutions support requirements for defining high availability for
continued operations with loss of primary data center. VMware virtualization software, with
recommendations and best practices for
Figure 8.28 shows disaster recovery providing acceptable levels of protection.
solution architecture with two data centers. The VMware virtualization software can be
Disaster recovery configurations ensure used to manage replication and
business operations continue following a synchronization of business software across
catastrophic data center failure. multiple data center locations.
Secondary data center must be designed • Virtual server replication to secondary

and maintained to support failover data center
operations. • Virtualization tools to orchestrate fail
over to a secondary data center
Custom IT administration processes must
be established to maintain and distribute Best practice: Use of virtualization
traffic between the data center locations. software technology to manage
distributed high availability systems
• Multiple site install and configuration supporting continued operations
following loss of the primary data center.

Business continuance: Server GIS components
Figure 8.29 ArcGIS Server components support high availability operations.
Figure 8.29 shows ArcGIS Server high machine (two or more GIS Servers) with
availability configuration. Server machines shared configuration store and server
that must be configured for high availability directories for highly available operations.
include the ArcGIS Server components and
supporting data sources. Warning: Third-party load balancers,
Web, and storage tier must be
This section will identify provisions for configured for highly available
supporting highly available ArcGIS Server operations.
tier components. Third-party vendor Best practice: Review architecture
solutions are available for building highly solution to ensure no single hardware/
available data sources hosted by DBMS network failure can cause failed operations.
and File share solutions.
Server GIS: Primary deployment patterns
Server GIS: Multi-machine architecture for high availability
Figure 8.31 ArcGIS Server high- availability solutions

can be supported with or without the ArcGIS Web
Adaptor.
Figure 8.30 ArcGIS Server site supports multiple Figure 8.31 shows options with and without
machine architecture. ArcGIS Web Adaptor. ArcGIS Server
production deployments may include two- or
Figure 8.30 shows ArcGIS Server multiple three-tier configurations.
machine architecture. ArcGIS Server site
architecture is designed to support a multi-

Two-tier Server configuration Two-tier Server configuration
• Highly available load balancer solution. • Highly available load balancer solution.
• Multi-machine ArcGIS Server site. • Multi-site identical load-balanced GIS
• Highly available file share for Server tier configuration.
Configuration Store and Server • Identical Configuration Store and Server
Directories. Directories for each site.
• Highly available data source (DBMS or • Highly available data source (DBMS or
File share). File share).
Three-tier Server configuration Warning: Virtual server deployments

must distribute platform tier components
• Highly available load balancer solution. on redundant host platforms.
• Multi-machine load balanced Web tier. Best practice: Review architecture
• Multi-machine ArcGIS Server site. solution to ensure no single hardware/
• Highly available file share for network failure can cause failed operations.
Configuration Store and Server
Directories. Server GIS: Pattern for disaster recovery
• Highly available data source (DBMS or
File share).
Warning: Virtual server deployments

on redundant host platforms.
Best practice: Review architecture
solution to ensure no single hardware/
network failure can cause failed operations.
Figure 8.33 Secondary data center must be designed
Server GIS: The Silo deployment pattern and maintained to support failover operations.
Figure 8.33 shows ArcGIS Server

architecture pattern for disaster recovery.
Primary high-availability solutions must be
established and maintained at each data
center.

Figure 8.32 ArcGIS Server Silo configurations require
solution to ensure no single
third-party configuration and load balancing solutions. hardware/network failure can cause
failed operations.
Figure 8.32 shows ArcGIS Server high
availability single machine site Custom business processes must be
configurations. ArcGIS Server single- established to maintain and distribute traffic
machine site configurations must support between the data center locations.
high availability with third-party solutions.

• Multiple site installs and configuration Best practice: System designed and
• Multiple site service deployments resources maintained to support
• Multiple site data source updates continued operations following loss of
• Concurrent application updates between primary data center.
sites
Business continuance: Portal for ArcGIS components
Figure 8.34 Portal for ArcGIS components support high-availability operations.
Figure 8.34 shows the Web GIS Portal for ArcGIS: Multi-machine
architecture components. Web GIS architecture
components that must be configured for
high availability include the Portal for
ArcGIS, ArcGIS Server, and ArcGIS Data
Store tier. The Portal for ArcGIS and ArcGIS
Data Store tier will be discussed first,
followed by a system-level discussion on
the Web GIS high availability and disaster
recovery solution.
Portal for ArcGIS components that must be

configured for high availability include the
Portal for ArcGIS components and Figure 8.35 Portal for ArcGIS can be configured with
supporting content store. This section will two Portal servers.
identify provisions for supporting highly
available Portal for ArcGIS tier components. Figure 8.35 shows a two machine Portal for
Third-party vendor solutions are available ArcGIS configuration.
for building highly available File share
solutions. Portal for ArcGIS architecture is designed to
support an active-active multi-machine (two
Portal server) configuration. Each portal
machine includes a system database and
index service. All inbound edit transactions

are redirected to the primary portal Figure 8.36 Portal for ArcGIS can support high-
machine, and primary database updates are availability solutions with or without the ArcGIS Web
Adaptor.
replicated to the secondary database. The
index service keeps users and item
Figure 8.36 shows high available portal
searches in sync between both machines.
architecture with or without using the
ArcGIS Web Adaptor. Portal for ArcGIS
Both portal machines store content in a
production deployments may include one-
shared content directory. The file-based
or two-tier configurations.
content directory must be configured on a
high-availability file share, and both portal
Single-tier Portal configuration (third party
machines must be configured with access to
HA load balancer)
the file share.
• Highly available load balancer
When one portal server machine fails, the
solution.
remaining machine will support all
• Two-machine Portal tier.
transactions.
• Highly available File share for
Content Store.
Warning: Third-party load balancers,
Web, and storage tier must be
Two-tier Portal configuration (ArcGIS Web
configured for high-availability
Adaptor with third party load balancer)
operations.
Best practice: Review architecture • Highly available load balancer
solution to ensure no single solution.
hardware/network failure can cause • Two-machine load balanced Web
failed operations. tier (include Web Adaptors).
• Two-machine Portal tier.
Portal for ArcGIS: Deployment patterns • Highly available File share for
for high availability Content Store.

solution to ensure no single hardware/
network failure can cause failed operations.

Business continuance: ArcGIS Data Store components
Figure 8.37 Portal for ArcGIS data store components.
Figure 8.37 shows the Portal for ArcGIS Portal deployments are supported by both
data store components. ArcGIS Data Store user-managed and ArcGIS-managed data
components must be configured for high stores.
Availability to support business continuance.
• User-managed data stores include data
This section will identify provisions for sources (Geodatabases and files).
supporting highly available ArcGIS Data
Store tier components. User-managed data store
• Geodatabases and file data sources for

Data Stores: User-managed vs. ArcGIS- ArcGIS Server published services.
managed • Database schema and content are
managed by the geodatabase
administrator.
• Geodatabase can be identified as a
managed data store for Portal hosted
sites.
ArcGIS managed data store
• Installed on Portal-hosted GIS Server

site.
• Stores feature data published to Portal
Figure 8.38 Data Store can be supported by for ArcGIS.
registered Geodatabase or the ArcGIS Data Store.
Best practice: ArcGIS Data Store

Figure 8.38 shows the available Portal Data
provides a scalable architecture for
Store options.
Portal for ArcGIS feature publishing.

ArcGIS Data Store: Is actually many
ArcGIS- managed Data Stores
Figure 8.39 ArcGIS Data Store supports storage of

different data types. Figure 8.40 ArcGIS Data Store supports a failover
Data Store instance.
Figure 8.39 shows the various ArcGIS-
managed Data Store options. Portal is Figure 8.40 shows the ArcGIS-managed
supported by several different ArcGIS- Data Store high availability configuration
managed Data Stores. option. Configure ArcGIS Data Store with a
standby machine so your feature data is
Supported feature datasets include the available even if the primary machine fails.
following:
Include a secondary failover machine with
• Features (points, polygons, lines) the Data Store install.
• 3D scene layers (tile cache)
• Real-time geoevent observations (high- • ArcGIS Server provides failover to
capacity "Big" Data Store) secondary machine.
• Common storage used by Primary
ArcGIS Data Store: Multi-machine and Secondary machines.
• ArcGIS Data Store automatically
architecture
generates backup recovery files.

Business continuance: Web GIS components
Figure 8.41 Web GIS supports high availability as an integrated designs solution.
Figure 8.41 shows an overview of

components supporting the Web GIS Web GIS components include Portal for
configuration. Web GIS components that ArcGIS, ArcGIS Server, and the ArcGIS
must be configured for high availability Data Store.
include the Portal for ArcGIS, ArcGIS
Server, and ArcGIS Data Store tier. Figure 8.42 shows the complete Web GIS
high-availability configuration.
All Web GIS components must stay in sync
to support high availability and disaster • High-availability Portal for ArcGIS tier
recovery solutions. (two Portal servers with shared Content
Store).
• High-availability ArcGIS Server tier
Web GIS: High-availability deployment (minimum of two GIS Servers with
pattern shared Configuration Store and Server
Directories).
• High-availability ArcGIS Data Store
(Primary Data Store with failover
secondary Data Store, sharing data on a
common file share) Data Store
automatically creates backups for
Figure 8.42 Web GIS high availability integrates disaster recovery.
Portal for ArcGIS, ArcGIS Server, and ArcGIS Data
Store solutions.

Web GIS: Disaster recovery deployment pattern
Figure 8.43 Web GIS provides a configuration backup and restore model for disaster recovery deployment.
Figure 8.43 shows the Web GIS disaster Custom IT processes must be established
recovery deployment pattern. Secondary to maintain and distribute traffic between the
data center must be designed and data center locations.
maintained to support failover operations.
• Multiple site installs and configuration
Primary high-availability solutions must be • Web GIS backup and restore model
established and maintained at each data provides replicated configuration
center. (available with 10.4 release).
• Backup and restore model provides Virtualization software technology can be

replicated configuration (available with used to manage replication and
10.4 release). synchronization of business software across
multiple data center locations.
must distribute platform tier components • Virtual server replication to secondary
on redundant host platforms. data center
Best practice: Review architecture • Virtualization tools to orchestrate fail
solution to ensure no single over to a secondary data center
hardware/network failure can cause
failed operations. Best practice: Use of virtualization
software technology to manage
distributed high availability systems
supporting continued operations
following loss of the primary data center.

Business continuance Business continuance
operations: Requires more operations: People and
than technology process considerations
Figure 8.44 Business continuance involves

technology, people, and process working together for
successful operations.
Figure 8.45 Business continuance solution must be

Figure 8.44 shows that business exercised to ensure compliance.
continuance involves more than technology
to ensure successful operations. Business Figure 8.45 shows best practices for
continuance operations are complex with ensuring business continuance. Providing
multiple components that must work the right people with the right skills in the
together to support reliable high availability right place at the right time is critical for
and disaster recovery solution. supporting sustained operations.
Technology solutions alone do not People qualifications
guarantee success.
• IT managed
• Qualified people must be trained to • Strong technical team
support recovery operations. • Knowledge of GIS and IT
• Processes must be implemented and
testing to ensure operational readiness.
Process validation
• Technology must work together to
support integrated operations
• Business alignment
• Established SLAs
• Knowledge management
• Training

Security strategy overview
Figure 8.46 Security is important to ensure effective business operations.
Figure 8.46 provides a summary of security Best practice: Finding the right balance
facts and recommended actions. Security is is important, and the right solution can
everybody's job, there is no exception. The be a moving target.
world is not a secure environment, and you
need to keep your eyes and minds open to
the threats around you.
Security resources
• Esri [Trust ArcGIS] security site
There is no single solution for security.
• [CSI Computer Crime and Security
Survey 2010-2011]
• There are costs and trade-offs that must
• [[https://www.nsslabs.com/ Web
be made to support an optimum
Browser Security Test Reports
solution.
• [Windows on Amazon EC2 Security
• Too much security controls can reduce
Guide]
productivity and increase cost.
• [Selected Documents on
• Too little attention and control can result
Confidentiality and Geospatial Data]
in loss of property and the ability to
• [SaaS, PaaS, and IaaS: A Security
perform.
Checklist]

GIS Product Architecture
Fall 2017 GIS Product Architecture 41st Edition
GIS Product Architecture shares the software components and platform configuration options
available for distributed GIS operations. Understanding application architecture alternatives and
associated configuration strategies provides a foundation for selecting an appropriate distributed
GIS design.
ArcGIS 10.5 is a major architecture release. An excellent overview of ArcGIS Enterprise is

provided in Philip Heede's presentation on Architecting your Deployment
ArcGIS technical architecture evolution
Figure 9.1 GIS architecture patterns have evolved from single- user stand-alone desktop- file-based systems to more
collaborative Web GIS systems of engagement.
Figure 9.1 shows how GIS architecture is geodatabase data source. Data was
evolving to enable a more adaptive and maintained and shared in an integrated
functional exchange of geographic database environment, improving
information. information continuity and quality of the
available data resources. Published data
File-based systems: Desktop applications could be managed and controlled to
building file-based datasets that were promote a common view of available
unique to the individual user. Building and validated data resources. Access to data
sharing information was limited to individual resources was limited to desktop users on
relationships, and data integration was the local area network.
limited.
Server-centric: Database resources were
Database-centric: Enterprise desktop published as Web services, making
clients would access a centrally shared information products available to a broad
GIS Product Architecture 9.1

Internet community of Web clients. Rich efforts to deploy new Web information
Internet clients could access services from products. Generic commercial applications
multiple server locations, expanding access able to leverage Web maps created and
and integration of information resources to a shared by business users provide timely
much broader community of users. access to information products at any
Applications were developed and deployed location on any supported device. Users
to leverage available Web services could create and administer their own
groups for sharing content, use configurable
Best practice: Database- and Server- apps to create new Web applications, and
centric architecture patterns support leverage solution templates to rapidly create
optimum governance for system of and deploy content to a broad community of
record content. users.
Web-centric: Introduction of a portal
architecture expanded development of Web Best practice: Web-centric portal
content to the business community, no architecture provides optimum solution
longer requiring software developer project for system of engagement.
ArcGIS system technical architecture
Figure 9.2 ArcGIS platform technical architecture includes Desktop, Server, Portal, a variety of ArcGIS Server roles,
and the associated data sources.
Enterprise-level GIS applications support a various ArcGIS architecture components is

variety of users throughout an organization, provided in Figure 9.2.
all requiring access to shared spatial and
attribute data sources. System hardware Central Data Servers: Shared spatial and
and software environments for distributed tabular database management systems
GIS applications are supported by a multi- provide central data repositories for shared
tier client/server or Web services geographic data. These database
architecture. A simple overview of the management systems can be located on

separate data servers or on the same GIS clients. Windows Terminal Servers host
central server platform. Data servers include GIS desktop applications on centrally
File Servers, File and Enterprise managed server farms allowing remote
Geodatabase Servers, Imagery servers, terminal clients to display and control
map and imagery tile cache, and database applications executed on the terminal server
servers supporting other Enterprise platforms. ArcGIS Server platforms support
business systems. Additional data a variety of Web applications and services
resources are available from ArcGIS Online accessed by standard browser clients, rich
and other Internet data sources. internet applications, or other desktop
applications.
ArcGIS Desktop Clients: ArcGIS Desktop
clients are licensed at three levels that Web GIS in the cloud: ArcGIS Online
address specific user application needs. Organizations provide GIS capabilities
Basic provides view and query functionality, through Esri-hosted software as a service
Standard provides functionality focusing on environment.
data maintenance and administration, and
Advanced provides the full suite of standard Web GIS on-premise: Base ArcGIS
desktop technology used by GIS Enterprise deployment (Portal for ArcGIS,
professionals. ArcGIS Engine and ArcGIS hosting ArcGIS Server site, ArcGIS Data
Runtime are programming environments for Store (Relational) supporting on-premise
developing custom ArcGIS Desktop system of engagement.
applications.
ArcGIS Enterprise server roles: ArcGIS
Hosted GIS services: GIS applications are Image Server (Raster Analytics and Image
supported within a distributed configuration Services), ArcGIS GeoEvent Server
by server platforms that execute GIS (realtime events), ArcGIS GeoAnalytics
functions. In a centralized solution, Server, and ArcGIS Business Analyst
Windows Terminal Server and ArcGIS Server—along with their associated ArcGIS
server platforms can host applications and Data Stores—provide expanded capabilities
services for a large number of concurrent for on-premise system of engagement.

ArcGIS Software Architecture
Figure 9.3 ArcGIS software component architecture
ArcGIS is an integrated collection of more detail in the Chapter 5 GIS Data

software components available for building Administration chapter.
a complete geographic information system.
The ArcGIS software products are used to ArcGIS Desktop can be deployed on client
deploy GIS functionality and business logic workstations or hosted by a Windows
where needed—in desktops, servers, Terminal Server. Custom ArcGIS Engine
custom applications, Web services, and and runtime applications include the same
mobile devices. The ArcGIS applications ArcGIS components supporting the ArcGIS
are supported by a common set of software Desktop commercial software and share
components. Figure 9.3 provides an common configuration strategies. Different
overview of the ArcGIS software component configuration alternatives are available to
architecture. support communications between the client
application and the GIS data source.
GIS is technology used for the creation, ArcGIS Desktop deployment patterns are
management, integration, analysis, display, discussed in more detail in the SDSwiki
and dissemination of spatial data. Spatial Chapter 2 Desktop operations section.
data includes any information that can be
associated with a location on the earth's ArcGIS Server is deployed in a scalable
surface or data that can be associated with Web application server architecture. The
a person or place that has a location. Web solutions include software
development kits and runtime environments
How the spatial data is maintained and supporting application development, system
published within the organization performance, and scalability.
contributes to the performance and Recommended platform configuration
scalability of the system design. GIS Data strategies will be provided for both standard
must be organized and managed to support and high availability architecture solutions.
effective and efficient GIS operations. GIS ArcGIS Server deployment patterns are
data deployment patterns are discussed in discussed in more detail in the [ SDSwiki
Chapter 2 Web operations].

GIS applications support an open systems (popular) software architecture solutions
architecture. The ArcGIS enterprise based on standard design practices cannot
architecture combines a variety of closely be overemphasized, since all parts of the
integrated commercial products to establish distributed configuration are critical and
a fully supported system solution. All must work together to ensure
commercial software products must be communication interfaces are properly
maintained to support evolving maintained and supported.
communication interface standards. The
importance of selecting well established
Virtualization deployment options

CPT Virtualization Model
Figure 9.4 There is more than one virtualization solution. Virtualization options include virtual sessions, virtual
servers, virtual desktops, and virtual client operations.
Figure 9.4 provides an overview of the four • Citrix: XenApp -> (Citrix Receiver)
principle virtualization solutions. • VMware: NA
Virtualization provides a variety of options
available for ArcGIS deployments. Note: CPT workflow: Citrix workflow with
Physical or Virtual platform tier selection.
Virtual session
Esri certifies each ArcGIS Desktop release
Virtual sessions are used for improved with Citrix XenApp server (Citrix Receiver)
security and high performance access to environment. A more complete discussion
shared centralized data souces. Multiple on Centralized Windows Terminal
desktop applications are deployed on a Server/Remote Desktop Services (Citrix)
centralized server platform sharing a Architecture is provided is provided later in
common server operating system. Desktop this chapter.
applications run on the server with each
session displayed and controlled by remote Warning: ArcGIS Pro is not supported in
terminal clients. a virtual session environment.
Best practice: Large number of Esri
Vendor offerings: customers use Citrix XenApp application
servers for remote user access to
• Windows: Remote Desktop Service centrally managed remote desktop
(RDS) -> Remote Desktop Connector (ArcMap) terminal services.
(RDC)

Virtual server Note: CPT workflow: Citrix workflow with
VDI platform tier selection.
Virtual Server machines reduce
administration overhead and share the host Esri customers have deploy Virtual Desktop
physical server compute resources. Virtual Infrastructure (VDI) and though they are
Server machines share a common known to work, Esri has not yet certified
hypervisor for access to host platform these solutions.
processing resources. Each Virtual Server
machine has its own installed Server Best practice: Virtual desktops
operating system. Server applications are environments are used by Esri for
installed in each separate Virtual Server development and hosted ReadyTech
machine. training environments.
Vendor offerings: Virtual desktop environments traditionally do

not have access to hardware video cards
• Windows: HyperV which has contributed to some functional
• Citrix: XenServer performance degradation noticed by users.
• VMware: VMware vSphere (ESX) Vendor solutions are available to include
shared NVIDIA GRID video cards within the
Note: CPT workflow: Virtual platform tier VDI host server architecture to improve
selection. display performance for future virtual
desktop deployment patterns.
Best practice: Large number of Esri
customers deploy ArcGIS Server in the
Best practice: Virtual desktops
three available virtual server
environments with host platform NVIDIA
environments.
GRID video cards are used for hosted
ArcGIS Pro deployments.
Virtual desktop infrastructure (VDI)
Virtual client
Virtual desktop infrastructured deployments
reduce administration overhead and share
Virtual clients are virtual desktop
the host physical server compute resources
environments deployed on a local physical
for centralized virtual desktop deployment.
desktop. Virtual Desktops interface through
Virtual Desktops share a common
a hypervisor for access to physical desktop
hypervisor for access to host platform
processing resources. Each Virtual Desktop
processing resources. Each Virtual Desktop
has its own installed Desktop operating
has its own installed Desktop operating
system. Desktop applications are installed
system. Desktop applications are installed
in each separate Virtual Desktop machine.
in each separate Virtual Desktop machine.
Virtual Desktops run on the client physical
Virtual Desktops run on the server with
workstation..
display and control provided by a variety of
remote terminal client devices.
Vendor offerings:
Vendor offerings:
• Windows: Virtual PC
• Citrix: XenDesktop
• Windows: Microsoft Virtual Desktop
• VMware: VMware VDI
Infrastructure
• Citrix: XenDesktop
• VMware: VMware View Note: CPT workflow: Increase client load to
handle visualization.

Best practice: Virtual desktops are used Warning: Some virtual desktop
by Esri for development and classroom environments to not have access to
training environments. hardware video cards which contributes
to some performance degradation
noticed by users.
ArcGIS Desktop architecture patterns
Figure 9.5 ArcGIS Desktop Software Architecture
ArcGIS Desktop software is supported on Best Practice: Communications between

Microsoft Windows desktop and terminal the ArcGIS Desktop application and an
server platform environments. QT Desktop Enterprise Geodatabase should be
Runtime provides ArcGIS components for limited to stable high-bandwidth local
custom desktop application development. network environments with minimum
Python is a fully supported scripting network latency.
language used for process automation.
Figure 9.5 provides an overview of the ArcGIS direct access to remote DBMS data
software components supporting the ArcGIS sources can result in reduced display
Desktop application. performance and unstable editing
operations. The most common architecture
ArcGIS Desktop is deployed in a for supporting remote client editing uses
client/server component architecture. The terminal access to a central Windows
client applications are tightly coupled when Terminal Server located in the SDE
accessing a DBMS data source, exchanging Geodatabase data center, which we will
hundreds of sequential data requests to discuss later in this chapter.
complete each user map display
transaction. A typical map display is Network latency is the primary cause for
refreshed in less than a second, requiring a reduced display performance when
very chatty protocol exchange with the accessing a remote database. Display
connected database. performance can be improved by reducing

the number of sequential application round map display. ArcGIS Desktop software
trip requests per display. You can reduce includes Web access to a variety of
the required round trips to the server by worldwide high quality ArcGIS Online
establishing a local feature cache or by basemaps and is fully integrated with
accessing remote DBMS data sources publishing and collaborating through ArcGIS
through an ArcGIS Server feature service. Online organization membership.
You could also establish ArcGIS Desktop
local basemap layers to further improve The CPT data source selections for an
display performance for a defined work ArcGIS Desktop workflow include a variety
area. These are some workaround options of data source formats (DB_DBMS,
for improving display performance, and may SFG_Small File GDB, LFG_Large File
not be practical for all workflow scenarios. GDB, SSF_Small Shape File, LSF_Large
Shape File, or Cache_Cached Tiles).
ArcGIS Desktop software connects to local Appropriate application query load
file and Imagery data sources, DBMS adjustments (Workstation/terminal server
sources, map and imagery cache, and Web desktop application or server SOC) are
data services. Web data services can be made based on the data source selection.
overlaid with local data in a standard GIS
ArcGIS Desktop workstation architecture
Figure 9.6 ArcGIS Desktop Workstation Architecture.
Four distributed ArcGIS Desktop The ArcGIS Desktop software will provide
workstation configuration patterns are native file access to GIS data located on
identified in Figure 9.6. These configuration local disk. GIS applications can access a
patterns include access to a networked file remote file data source by using Microsoft's
data source, direct connect access to an Common Internet File Services (CIFS) or
Enterprise Geodatabase, direct access to similar UNIX Network File Services (NFS).
a supported DBMS (non-SDE), and DBMS When mounting the remote disk, the remote
access through an ArcGIS Server feature file would appear as a local file share to the
service. desktop application. Query processing for a

file data source is supported by the ArcGIS service. You can then make edits to the
Desktop application. local copy in ArcMap and synchronize the
edits back to the service. Edits can be made
ArcGIS Desktop software provides a direct to the local copy without having to be
connection to all supported database connected to the server. Access to the
servers. The direct connect option (Direct server is only required when creating the
Connect) includes the direct connect API local copy or applying changes from the
executables and will communicate with a local copy to the server. This workflow can
database client installed on the same be useful when your organization has
machine. The database client will support disconnected employees and provides a
network communications to the database common method for editing the same data
server. using multiple clients, such as through the
web or using desktop applications. The
ArcGIS Desktop software provides direct functionality is built into ArcMap and does
connections to supported database not require any customizations. Edits to a
servers providing view, query and analysis published feature service are captured in a
of the DBMS data content. Some of the single version of the database.
databases you access can contain
geodatabase tables, functions, and
procedures, but they don't have to; you can Four ArcGIS Desktop workstation
connect to any supported database and configuration alternatives are identified in
view the data from ArcGIS Desktop. Figure 9.6. All configurations are supported
by ArcGIS Desktop applications installed on
Enterprise geodatabases, also known as a local workstation. ArcGIS Desktop
multiuser geodatabases, are stored in a workstation configuration options include
relational database using Oracle, Microsoft access to a network file data source, direct
SQL Server, IBM DB2, IBM Informix, or connect access to an Enterprise
PostgreSQL. These geodatabases require geodatabase, direct access to a non-SDE
the install of ArcSDE schema and can be DBMS, and DBMS access through an
unlimited in size and numbers of users. ArcGIS Server feature service.
ArcMap allows you to edit a supported CPT Calculator ArcGIS Desktop

database by creating a local copy of data workstation configurations
from a published ArcGIS Server feature

Centralized Windows Terminal Server/Remote Desktop Services (Citrix)
Architecture
The Microsoft Windows Terminal Server (name changed to Remote Desktop Server <RDS>
with Windows Server 2008 R2 release) operating system establishes a multiuser environment
on a Windows server host. A Windows terminal client (name changed to Remote Desktop
Connection with Windows Server 2008 R2 release) provides display and control of applications
executed on the Windows Terminal Server. Microsoft uses a standard Remote Desktop Protocol
(RDP) for communication between the terminal server and the Windows client. Windows
terminal server platform memory recommendations are generated based on peak
concurrent ArcGIS Desktop user sessions supported by the selected platform configuration.
The Citrix Xen Application Server (XenApp) XenApp provides many additional benefits
enables a more efficient independent over just RDS alone, including “seamless”
computing architecture (ICA) protocol to windows, universal print drivers, and HDX
communicate between the terminal server technologies such as HDX 3D Progressive
and client Windows platform. The ICA Display for imagery acceleration, just to
protocol requires less than 28 Kbps name a few. Most Esri customers that
bandwidth (rendering vector data deploy centralized thin-client solutions have
information products) for full Windows realized the benefits of Citrix XenApp and
display and control of GIS applications have deployed it in addition to standard
supported on a Windows Terminal Server. RDS. There is currently a very large Esri
Traffic can increase to 100 Kbps bandwidth customer user base that utilizes Citrix
when accessing a raster data source. XenApp for ArcGIS Desktop ArcMap
XenApp supports client software for operations. The following knowledge base
Windows, UNIX, Macintosh, and embedded article provides Esri best practices for
Web client applications. running ArcGIS Desktop ArcMap in a
Citrix XenApp environment.
Figure 9.7 Centralized Remote Desktop Client architecture

Esri initially certified ArcGIS 10 SP2 as a to a network file data source, direct connect
hosted application with Citrix XenApp 6 and access to an Enterprise geodatabase, direct
Windows 2008 R2 using the Citrix ICA access to a non-SDE DBMS, and DBMS
Online Plug-in for Windows 12.1. Esri has access through an ArcGIS Server feature
since certified ArcGIS 10.3 as a hosted service.
application with Citrix XenApp 6 and
XenApp 6.5 with Windows 2008 R2 using Terminal clients have a persistent
the Citrix ICA Online Plug-in for Windows connection with the Windows Terminal
13.1. Server ArcGIS Desktop session; lost
connections are reinstated without losing
Though XenApp is a popular solution and the session (referred to as Session
meets most desktop application publishing Reliability in XenApp). The application
needs, there are a few limitations with its display is provided over the network to the
use in GIS. The limits are primarily centered terminal client, requiring much less data
on supporting graphic intensive displays transfer than the spatial data query chatter
such as those from ArcGlobe and other 3D between the application and the data
applications. These limits can result in the source. The terminal client display traffic
need to maintain some presence of thick- requirements are very small; supporting
client desktops for targeted power users good application performance over 28 Kbps
that require high-performance and 3D modem dial-up connections (displays with
graphics environments. Design and testing an image backdrop require more
efforts are in work with our vendor partners bandwidth).
to improve 3D performance in remote
desktop deployment patterns. ArcGIS Each ArcGIS Desktop user session hosted
Resources blog shares Esri test feedback on the Windows Terminal Server connects
on ArcGIS Pro with NVIDIA K1 in to each data source the same way as a
XenDesktop. Citrix shares results for their distributed client workstation session. Most
commitment with Esri in supporting ArcGIS current Esri customers using Windows
Pro for Citrix XenApp and XenDesktop. Terminal Server also use the Citrix XenApp
Server software. The Windows Terminal
Further, centralizing applications and data Server farm is supported by commodity
naturally creates a distributed printing Windows server platforms (Intel or AMD).
environment. This can be particularly Client session load balancing across the
problematic for large plot printing used in terminal server farm is managed by the
GIS. Solutions are available to help with Citrix software. Client profiles and security
remote printing management and these options are provided to support an optimum
solutions are being used by Esri Citrix GIS user display experience.
customers. The following reference
document addresses printing with ArcGIS
in Citrix environments and mentions some CPT Calculator remote ArcGIS Desktop
of the available solutions. Two of the most (ArcMap and ArcGIS Pro) configuration.
popular have been ThinPrint and UniPrint.
ArcGIS Pro reference sites
Four distributed ArcGIS Desktop Citrix (recommendations for ArcGIS Pro VDI
configuration alternatives are identified in host platform configurations)
Figure 9.7. All configurations are supported
by remote terminal client access to ArcGIS • Virtualizing ArcGIS Pro: Nvidia Grid
Desktop applications hosted on a central vGPU Profiles May 2015 Deployment
Windows Terminal Server. ArcGIS Desktop configuration: 10-12 ArcGIS Pro Virtual
Citrix configuration options include access Desktops on high performance 24-core

server with two NVIDIA GRID video Best Practice: Provide dedicated host
cards. platform environment for ArcGIS Pro
• Virtualizing ArcGIS Pro: Nvidia GRID virtual desktop deployment.
Tesla M60 July 2017 Recommended ArcGIS Pro host platform
• ArcGIS Desktop Virtualization Appliance solution
March 2017 Deployment configuration:
Up to 25 ArcGIS Pro Virtual Desktops
on high performance 28-core server with
NVIDIA Tesla GPU accelerator cards.
ArcGIS Enterprise services architecture
Figure 9.8 ArcGIS Enterprise software licensing
Figure 9.8 shows the ArcGIS Enterprise • An optional web adaptor component is
software licensing components. ArcGIS included for enhanced security and
Enterprise licensing includes software for a network load balancing.
basic deployment along with options for • Portal for ArcGIS software provides
additional separate ArcGIS Server roles. overall content management for
information security, collaboration, and
ArcGIS Enterprise software components sharing.
include Portal for ArcGIS, ArcGIS Server, • ArcGIS Server roles expand capabilities
ArcGIS Web Adaptor, and five additional to include ArcGIS Image Server, ArcGIS
server roles. GeoEvent Server, ArcGIS GeoAnalytics
Server, and ArcGIS Business Analyst
ArcGIS Enterprise software licensing Server.
• ArcGIS Server is delivered as a single ArcGIS Server roles are designed for rapid
software install that includes web deployment, user collaboration sharing, and
service endpoints and SOC functions friendly administration.
within a single software bundle.

ArcGIS Enterprise deployment capabilities.
strategies
ArcGIS Server software architecture

Ports used by ArcGIS Server
Web mapping services provide an efficient

and effective approach to serving map
products and services over the Internet. The
ArcGIS Desktop architecture presented
earlier in this section requires tightly
coupled client/server processes that
Figure 9.9 ArcGIS Enterprise deployment strategies demand stable high-bandwidth
communications supported over relatively
Figure 9.9 shows the ArcGIS Enterprise short distances. Web client communications
deployment strategies. ArcGIS Enterprise are supported using a transaction-based
provides what you need to enable HTTP, which supports optimum
enterprise GIS content development and communications over long distances and
sharing based on your business needs. less stable communication environments.
ArcGIS Enterprise provides a component-

based architecture that can be deployed
and scaled as required to satisfy GIS
operational needs.
ArcGIS Enterprise base deployment
• Relational data store supports feature

data tables. Figure 9.10 ArcGIS Server Software Architecture
• Portal for ArcGIS provides content
management and security. ArcGIS Server services are published
• Hosting ArcGIS Server manages Portal through a Web server. Web server clients
published services. are presented with a catalog of published
• ArcGIS Data Stores manage Portal services when accessing the Web site. Web
published data resources. applications consume map services and
render a client presentation layer to support
Server roles the published application workflow. Client
and Web servers are loosely coupled
• ArcGIS Server supports registered GIS (stateless) connection in which each client
services. communication represents a complete
• ArcGIS Image Server supports image transaction. Transactions are processed by
services and raster analytics. the appropriate Web-based GIS server and
• ArcGIS GeoEvent Server supports real- returned to the client.
time events.
• ArcGIS GeoAnalytics Server supports Esri Web GIS services are hosted by
spatial and temporal analysis. ArcGIS Server software. ArcGIS Server
provides an ArcGIS software-based
Base ArcGIS Enterprise deployment is environment for deploying GIS server-based
required to support ArcGIS Server role ArcGIS applications and services. ArcGIS

Server can be deployed as a Web service integrated GIS Server software bundle. An
for Internet/Intranet clients or as a network optional Web Adaptor component is
service for local desktop clients. Rich included with the ArcGIS Server install
internet applications (RIA) Web clients are package for improved security and network
loosely connected, lightweight, handheld or load balancing.
desktop computers that can support a
variety of Web based applications and
include persistent data cache and ArcGIS Server key site aware
potentially support disconnected GIS client component functions
operations. Client application deployment
and data exchange are managed by ArcGIS
Server parent services. ArcGIS Imagery
services are fully integrated into ArcGIS
Desktop and ArcGIS Server with the ArcGIS
10 release.
The software architecture components for

ArcGIS Server include Web browsers, RIA
clients, GIS Server, and geodatabase
software components that can be deployed
on different platform combinations to
support scalable capacity and system
availability requirements. Location of the Figure 9.11 ArcGIS Server architecture terminology
various software components and the
selected software configuration contributes Figure 9.11 provides an overview of the
to system capacity, service reliability, terminology used to describe the ArcGIS
security, and overall output performance. Server architecture. An ArcGIS Server high
availability configuration will be used to
ArcGIS 10.1 was a major release for Server. introduce these terms. These terms will be
ArcGIS 10.1 for Server introduced all the discussed in more detail as we describe the
primary functionality of ArcGIS 10 delivered ArcGIS Server architecture patterns later in
in a new 64-bit JAVA based software this chapter.
component architecture. ArcGIS 10.1 for
Server was designed for rapid deployment The initial ArcGIS Server install defines the
and friendly administration. Server primary components that make up a named
enhancements improved installation, Site. These primary components include the
performance, reliability, administration, GIS Server and the Site Configuration Store
cloud deployment options, and more Linux (ConfigStore). The Site ConfigStore
friendly software compatibility. Primary contains all service definitions and
software components associated with the configuration parameters for GIS Server
ArcGIS Server software architecture are machines within the named Site. Data
identified in Figure 9.10. GIS Server support source path names are included as part of
access to file data sources, ArcSDE each service definition. Each GIS Server
Geodatabase and non-geodatabase includes HTTP end points (SOAP, REST,
database sources, imagery and cached tile Open Standards) available for publishing
data sources. and serving Web services. HTTP access to
GIS Server services is through HTTP port
ArcGIS Server is delivered as a single 6080. HTTPs secure communications with
software install that includes the web GIS Server is provided through port 6443.
service endpoints within a new fully

When you join a new GIS Server machine Each GIS Server machine can participate in
to a named Site, it will automatically deploy only one Site Cluster. The Site ConfigStore
service configurations as defined in the Site will contain service configurations for all
ConfigStore (service configurations are Clusters defined within the named Site. The
defined per GIS Server machine). The Site ConfigStore file share will also include
ConfigStore contains a single repository that a Server Directories file for collecting and
defines all GIS Server machine sharing common administrative Site
configurations within a named GIS Server directories.
Site. All GIS Server machines within a site
will share a single ConfigStore (a replicated ArcGIS 10.3 introduced the option for
copy of the ConfigStore would create a new single-cluster configuration, with service
Site). ConfigStore must be located on a file handler load balancing disabled. Single-
share which can be accessed by all GIS cluster deployment provides more scalable
Servers within the named Site. architecture and is the baseline for the
ArcGIS 10.4 and 10.5 deployments.
Warning: If you lose the ConfigStore or
data source, you lose the site. Inbound service requests will be assigned to
ConfigStore backup and recovery was one of the GIS Servers located within the
introduced with the 10.2 release.. named Site. Inbound requests will be
distributed across the site machines. In a
ArcGIS Server includes optional Web single-cluster site, the machine that
Adaptor software that can be used to receives the inbound requires will complete
manage GIS Server inbound service the service transaction.
requests. The Web Adaptor must be
installed on a machine with a third party In a multi-cluster site configuration, a GIS
Web Server. The Web Adaptor is Site Server service handler will assign the
aware, which means that it will distribute service request to the first available service
inbound service requests across the active instance within the named Site. If a service
GIS Server machines within the named Site. instance is not available on the assigned
The Web Adaptor can also act as a reverse GIS Server machine, the service handler
proxy, accepting service requests on a will send the request to an available service
configurable inbound port (i.e. Port 80) and instance on another one of the GIS Server
communicating with the Site GIS Server machines within the named Site for
machines on Port 6080. If any of the Site processing. All GIS Server service handlers
GIS Server machines fail, the Web Adaptor are Site aware, which means they are able
will route service requests only to the to assign an inbound service to any
remaining active GIS Server machines (high available service instance on any GIS
availability failover functionality). A single Server machine within the named Site
Web Adaptor can service only one named (service handler load balancing).
Site. Multiple Web Adaptors can be
assigned to a single named Site. Multiple High availability implies there is no single
Web Adaptors can be deployed on a single software or server component failure that
Web server (enables support for multiple will results in a Site failure. A high
ArcGIS Server named Sites from a single availability configuration must include
Web server). network load balancing with failover
capability to ensure inbound traffic will be
A GIS Server named Site can include one routed to an active Web gateway machine,
or more GIS Server Clusters. All GIS Server at least two GIS Server machines, and high
machines within a defined Site Cluster will availability File Share for the ConfigStore,
deploy the same service configurations. Server Directories, and any shared File data

sources. Any DBMS data source would also ArcGIS platform configuration strategies
need to be supported in a failover
configuration. If you lose the ConfigStore or • Software as a Service (SaaS)
any of the required data sources, you lose o ArcGIS Online Organization;
the Site. Hosted services online; Online
basemaps and services
ArcGIS Server uses certain default ports to • Software/data installed on-premises or
communicate with machines over the in a private cloud
network. The Ports used by ArcGIS 10.2 for o Portal for ArcGIS; ArcGIS
Server are identified in the Esri ArcGIS Server; Data Appliance for
Help. ArcGIS
Best Practice: Deployment strategies

ArcGIS Platform can include a mix of self-managed and
vendor-managed hardware option.
deployment strategies
There are a few self-managed deployment
options:
• Non-cloud on-premise ArcGIS Server

deployment
• Infrastructure as a Service (IaaS)-based
community private cloud deployment
• Hybrid deployment, including IaaS and
on-premises services
Figure 9.12 ArcGIS Platform deployment strategies There are also vendor-managed
include a mix of Cloud based SaaS and on-Premise
hosting options.
deployment options:
• IaaS-based community private cloud

ArcGIS platform deployment strategies can
include a mix of self-managed and vendor- deployment
• Hybrid of IaaS public and private cloud
managed configuration options.
services
• Vendor-managed IaaS-based public
Key ArcGIS components for Web GIS
cloud deployment
• Public SaaS-based deployment
• Web Apps
• GIS portal
How you deploy your ArcGIS Platform
• GIS service framework
depends on your business requirements, as
• Ready-to-use content
described by this video on ArcGIS
Deployment Patterns.

Portal for ArcGIS platform configuration
Figure 9.13 Portal for ArcGIS Platform Configurations can include registered Web services, federated ArcGIS Server
sites, and a hosting ArcGIS Server site.
Figure 9.13 shows an overview of the Portal detail in the Information Security chapter.
for ArcGIS configuration. Portal for ArcGIS Named users of the Portal organization can
can be installed on a stand-alone web create Web maps, add services to their
server or as a content management content, and share content with groups and
component of a federated ArcGIS Server other members throughout their
configuration. organization.
Portal for ArcGIS enables secure and Portal for ArcGIS releases starting with 10.3
private content sharing within the include documentation for installing a high
organization and leverages mobile, server, availability Portal configuration.
and desktop clients.
Registered Web services
Portal for ArcGIS is installed on a Web
server with a dedicated ArcGIS Server Web Web services referenced in Web maps or
adaptor. The server includes an identity added to the Portal are considered
store which contains Portal member user registered services. Web maps can be
names, passwords, and roles. Portal for created from Public or internal published
ArcGIS security authentication and Web services.
authorization options are discussed in more

Federated ArcGIS Server sites capabilities and deployed GIS Server roles
drive ArcGIS Data Store selection.
Internal ArcGIS Server sites can be
federated with Portal for ArcGIS. Federated Portal for ArcGIS can be deployed in a
ArcGIS Server site web services are variety of configurations adapting to your
published and shared by the Portal content hosting requirements. Portal for ArcGIS is
management interface. Federated ArcGIS supported by a dedicated Web Adaptor.
Server sites are managed as part of the
Portal configuration. Access authorization Best Practice: Use the Portal for ArcGIS
for all federated ArcGIS Server sites are Data Store (relational, introduced with
managed by the Portal for ArcGIS identify ArcGIS 10.3) for scalable feature service
configuration. deployment.
ArcGIS 10.4 introduced capabilities for fine The Capacity planning tool includes
grained access control of federated ArcGIS features for sizing a Portal for ArcGIS
Server sites. You can update a federated configuration.
server site to restrict publishing and
administrative access. Once updated, all
portal publisher and administrator access ArcGIS Enterprise platform
will be controlled by group privileges unique configuration structure
to the restricted federated server.
Note: Only ArcGIS Server sites using

version 10.2 or later can be federated with a
portal.
Hosting server site
A federated ArcGIS Server site can be fully

integrated with your portal if you designate it
as a hosting server. A hosting server allows
portal users to:
Figure 9.14 ArcGIS Enterprise platform tier
• Publish tiled map and feature services architecture
to the portal from ArcGIS Desktop.
• Share layers and maps from Esri Maps Figure 9.14 shows the ArcGIS Enterprise
for Office. base deployment and options for deploying
• Create maps by adding CSV files and the ArcGIS Server roles. ArcGIS Enterprise
shapefiles from local machines to the architecture includes single-tier, two-tier,
portal map viewer. and three-tier platform configuration options
• Publish CSV and shapefiles as feature for adaptable data center management.
services from the portal website.
ArcGIS Enterprise is designed to support
ArcGIS Data Stores multiple single-, two-, and three-tier platform
architecture solutions. All of the optional
ArcGIS Data Stores require a hosting ArcGIS Server site deployment options are
ArcGIS Server. The ArcGIS Data Stores available for the ArcGIS Enterprise base
manage data for services published through deployment and for each of the ArcGIS
Portal for ArcGIS. Portal publishing Server roles. The optimum architecture
solution is established based on operations

security, systems management • The ArcGIS Server tier can scale out to
(environment isolation), and system satisfy availability and capacity
capacity requirements. requirements.
• The DBMS tier can include a failover
Platform tier structure platform to satisfy high availability
requirements.
• Base deployment: Includes Portal for
ArcGIS, ArcGIS Server (hosted), and ArcGIS Server is designed to support a
required ArcGIS Data Stores. scalable Web architecture. Optimum
• Single-tier structure: All software platform environments are configured using
components are deployed on a single standard physical or virtual server platform
platform tier. technology. ArcGIS Server is licensed
• Two-tier structure: Software based on the number of platform processor
components are deployed on two core (virtual or physical) supporting the GIS
separate platform tiers. Server software tier.
• Three-tier structure: Software
components are deployed on three Best Practice: Each platform tier can
separate platform tiers. support multiple servers to satisfy
capacity requirements.
High-availability scalable architecture
patterns
• Both web tiers can scale out to satisfy

availability requirements.
ArcGIS Server site single-tier platform configuration
Figure 9.15 Single-Tier Platform Configurations
The ArcGIS Server installation takes less Figure 9.15 provides an overview of single-
than 5 minutes, and a single machine is tier platform configurations. Single-tier
ready for publishing services from ArcGIS configurations provide one or two platforms
Desktop without any additional installation capable of supporting all Web service
or configuration requirements. components.

Minimum Configuration: A complete Web Server machines throughout the Site to
site can be supported on a single hardware avoid having requests back up on one
platform. This configuration is appropriate server when extra processing resources are
for Web service development and testing, available on the other server (load
Sites with a limited number of service balancing is automatically handled within
requests, and initial prototype deployments. the GIS Server Site), (3) common high
A special single chip (2-core) workgroup availability (fault tolerant) file share for the
server license bundled with a Microsoft SQL ConfigStore and SvrDirectory files, and (4)
Server database is available for customer replicated file and DBMS data sources with
sites that can be supported by a single local copy distributed on each GIS Server
platform configuration. machine (in some cases, performance can
be adequate with file data located on the HA
If more than one GIS Server machine is file share).
joined to a minimum configuration with
inbound service requests all sent to the The GIS Servers can publish services
initial GIS Server machine, the initial GIS through their native Port 6080, or a third
Server service manager will assign inbound party Web server with the Web Adaptor can
service requests to any available GIS be installed on the ArcGIS Server tier for
Server machine instances within the named enhanced security.
Site. The ConfigStore and SvrDirectories
must be shared with all GIS Server For Imagery and File Geodatabase data
machines within the named Site and the sources, data should be deployed on each
GIS Data source must be distributed to GIS Server machine or configuration should
each machine to avoid performance ensure high bandwidth dedicated storage
contention. network connection and high performance
storage architecture to avoid potential heavy
This is not a high-availability configuration, traffic and I/O contention (network or
since if you lose the initial GIS Server storage contention can contribute to poor
machine you will lose the ConfigStore and Site performance).
service request distribution to the remaining
GIS Server machines within the Site. In a mixed services environment, it is good
practice to deploy image services and
High-Availability GIS Server Site Cached tile services on a separate GIS
Configuration: Most GIS server production Server machine from dynamic mapping
operations require redundant server services.
solutions, configured so the site remains
operational in the event of a single platform CPT Calculator tab can be used to
failure. A high-availability configuration must configure and complete a single-tier System
continue to support production operations Architecture Design capacity planning
during single platform maintenance and analysis.
upgrade and while configuring and
publishing new services. This configuration CPT Calculator ArcGIS Server single-tier platform
includes (1) network load balancing to route configuration
the traffic to each of the ArcGIS Server
machines during normal operations and
only to the active server if one of the servers
fails, (2) GIS Server Site aware service
manager load balancing to distribute spatial
services processing load between the GIS

ArcGIS Server site two-tier platform configuration
Figure 9.16 Two-Tier Platform Configurations (Separate Data Servers)
ArcGIS Server site architecture is designed High-Availability Configuration: Most GIS

for easy install and administration. The two- server production operations require
tier architecture in Figure 9.16 includes redundant server solutions, configured so
separate GIS server and data server the site remains operational in the event of
platform tier. The Web server and GIS a single platform failure. This configuration
server components are located on the GIS will continue to support production
server platform tier, and the data server is operations during single platform
located on a separate data server platform. maintenance and upgrade and while
This is a popular configuration for sites with configuring and publishing new services.
large volumes of data resources or existing This configuration includes (1) network load
data servers. A single copy of the data can balancing to route the traffic to each of the
support multiple GIS Server machines in ArcGIS Server machines during normal
conjunction with other enterprise GIS data operations and only to the active server if
clients. one of the servers fails, (2) GIS Server Site
aware service manager load balancing to
For Imagery and File Geodatabase data distribute spatial services processing load
sources, data should be deployed on each between the GIS Server machines
GIS Server machine or configuration should throughout the Site to avoid having requests
ensure high bandwidth dedicated storage back up on one server when extra
network connection and high performance processing resources are available on the
storage architecture to avoid potential heavy other server (load balancing is automatically
traffic and I/O contention (network or handled within the GIS Server Site), (3)
storage contention can contribute to poor common high availability (fault tolerant) file
Site performance). share for the ConfigStore and SvrDirectory

files, and (4) two DBMS servers that are In a mixed services environment, it is good
clustered and connected to a common practice to deploy image services and
storage array data source. The primary data cached tile services on a separate GIS
server supports query services during Server machine from dynamic mapping
normal operations, and the secondary data services.
server takes over query services when the
primary server fails. Data server clustering CPT Calculator tab can be used to
is not required if availability requirements configure and complete a two-tier System
are satisfied with a single data server. Architecture Design capacity planning
analysis.
The GIS Servers can publish services
through their native Port 6080, or a third- CPT Calculator ArcGIS Server two-tier platform
party Web server with the Web Adaptor can configuration
be installed on the ArcGIS Server tier for
enhanced security.
ArcGIS Server site three-tier platform configuration
Figure 9.17 ArcGIS Server Three-Tier Platform Configurations
Three-tier configurations include Web reverse proxy and network load balancing
server, GIS Server, and data server tiers. on the Web Server tier, and would likely be
the most popular solution. The three-tier
Figure 9.17 shows an ArcGIS Server three- configuration provides a scalable
tier configuration. This configuration architecture, where the middle tier can
includes the Web Adaptor which provides support two or more platforms as required
to support capacity requirements.

High-Availability Configuration: server fails. Data server clustering is not
required if availability requirements are
Most GIS server production operations satisfied with a single data server.
require redundant server solutions,
configured so the site remains operational in The GIS Servers can publish services
the event of a single platform failure. This through their native Port 6080, or a third
configuration will continue to support party Web server with the Web Adaptor can
production operations during single platform be installed on the Web Server tier for
maintenance and upgrade and while enhanced security (reverse proxy server)
configuring and publishing new services. and site aware load balancing.
This configuration includes (1) network load
balancing to route the traffic to each of the Warning: For Imagery and File
Web Servers during normal operations and Geodatabase data sources, data should
only to the active service if one of the server be deployed on each GIS Server machine
machines fail, (2) network load balancing to or configuration should ensure high
distribute the traffic to each of the GIS bandwidth dedicated storage network
Server machines during normal operations connection and high performance
and only to the active servers if one of the storage architecture to avoid potential
servers fail, (2) GIS Server Site aware heavy traffic and I/O contention (network
service manager load balancing to distribute or storage contention can contribute to
spatial services processing load between poor Site performance).
the GIS Server machines throughout the
Best Practice: In a mixed services
Site to avoid having requests back up on
environment, it is good practice to
one server when extra processing
deploy image services on a separate GIS
resources are available on the other server
Server machine from dynamic mapping
(load balancing is automatically handled
services.
within the GIS Server Site), (3) common
high availability (fault tolerant) file share for
the ConfigStore and SvrDirectory files, and CPT Calculator tab can be used to
(4) two DBMS servers that are clustered configure and complete a three-tier System
and connected to a common storage array Architecture Design capacity planning
data source. The primary data server analysis.
supports query services during normal
CPT Calculator ArcGIS Server three-tier platform
operations, and the secondary data server configuration
takes over query services when the primary

Data center platform tier architecture
Figure 9.18 ArcGIS Enterprise data center platform tier architecture
ArcGIS Enterprise includes an expanding Best Practice: High availability requires

number of servers. Figure 9.18 shows the platform redundancy.
variety of platforms supporting the Web GIS Warning: All components of the system
product architecture. must be configured with sufficient
capacity to support peak business
Enterprise solutions support four generic needs.
platform tiers.
Best Practice: Capacity planning is
• WTS tier: Centralized ArcGIS Desktop essential to ensure deployment success.
deployment
• DBMS tier: Variety of database CPT Design ArcGIS Enterprise system
management solutions architecture design with ArcGIS Desktop
• GIS tier: Variety of GIS application (ArcMap and ArcGIS Pro) deployments.
servers
• Web tier: ArcGIS Web Adaptors and Generic ArcMap ArcGIS Enterprise
Portal servers. platform technical architecture
Generic Pro ArcGIS Enterprise
Separate platform environments (Web, platform technical architecture
Portal, Servers, Data source)may be Detailed Pro VDI ArcGIS Enterprise
required for internal and external services to platform technical architecture
satisfy security requirements.

ArcGIS Server site deployment (single-site alternative
patterns)
Figure 9.19 ArcGIS Server site configuration options
Figure 9.19 shows two alternative ArcGIS Warning: Multiple-cluster site scalability
Server site configuration options. ArcGIS limited for light services with high
Server alternative site configurations are transaction rates.
available since the ArcGIS 10.4 release.
ArcGIS Server read-only operations

ArcGIS Server site
deployment (multiple site
• Cached copy of ConfigStore and patterns)
SvrDirectories is maintained on each
machine within the ArcGIS Server site.
• Site is restricted to read-only operations.
Best Practice: Read-only configuration

provides sustained operations with loss
of network connection with shared site
configuration store.
ArcGIS Server multiple-cluster site

operations
Figure 9.20 ArcGIS Server multiple-site configurations
can be deployed to satisfy more advanced data
• Load balancing provided by Web center deployment patterns.
Adaptor or third-party load balancer
• Additional service manager load Figure 9.20 shows the ArcGIS Server
balancing within site. multiple-site deployment patterns. ArcGIS
Server can be deployed as individual
machines in a multi-site deployment
architecture, with each machine having its

own dedicated ConfigStore and ArcGIS Server can be deployed as multiple
SvrDirectories. This architecture pattern, independent GIS Server site machines as
although more complex to administer, can shown in Figure 9.21. Independent GIS
provide advantages for some specific Server site configurations remove the
business workflows. deployment simplicity built into the adaptive
GIS Server site architecture.
Some multiple-site examples include the
following: Business requirements drive how these
sites would share service requests and
• Separate development, staging, and common data resources. Multiple GIS
production server environments. Server site configurations are not aware of
Recommended practice for managing each other, and proper business processes
production architecture change. must be established for keeping data,
• Failover ArcGIS Server role active- security, and services in sync. Third party
passive configurations. Licensing load balancer, often an integral part of many
required only for active ArcGIS Server. large data center Web operations, must be
• Separate replicated data center used to distribute inbound traffic to the
configurations. Support continuity of independent GIS Server sites.
operations planning (COOP) and
disaster recovery (DR) configurations. Custom business processes must be
• Replicated scale-out site configurations. established to maintain and distribute traffic
Optimized high capacity virtual GIS across the GIS Server multiple site
server deployments. configuration.
Single-tier, two-tier, and three-tier • Multiple site install and configuration.

architecture deployment strategies Each independent server machine site
discussed discussed earlier in this section must be installed and configured. Any
can be considered for each of the multiple- required load balancing and failover
site configuration patterns. requirements must be satisfied by a
third-party solution (Web Adaptor does
not support multiple site load balancing
ArcGIS Server multiple site operations). In a virtual server
integration environment, machine clones and site
management tools can be used to
deploy single tier GIS Server machine
sites.
• Multiple site service deployments.

Custom business processes must be
established for publishing and updating
services across the multiple server
machine farm. Published service
definitions (SDs) can be used to
distribute services across multiple GIS
Server site machines. Properly
registered data folders and database
Figure 9.21 ArcGIS Server multiple Site deployment sources can simplify service deployment
platforms are not aware of each other. Administrator configuration requirements. Custom
is responsible for deploying and maintaining the ArcPy scripts may be used to automate
multiple site configuration. service deployments.

• Multiple site data source updates. • Production – This site supports live
Data sources for all server machines business workflows. Only changes that
must be in sync. Shared data sources have passed testing on the staging site
can be used for two and three tier Web are moved to the production site.
architecture patterns when server
machines are located in a common Guidelines and best practices for building
location. For multiple data center and/or and managing ArcGIS Server in
single tier configurations, data must be development, staging, and production
replicated or replaced on each GIS environments are available in the reference
Server data source location. ArcGIS Help documentation.
• Updating applications between sites.

Custom business processes can be Active-Passive ArcGIS Server Failover
established to copy application files and configurations
update Web URLs across each GIS
Server site.
Enterprise development, staging, and

production operations
Figure 9.23 Active-passive failover

configurations can provide high availability
with minimum impact on licensing costs.
Figure 9.22 Enterprise operations require
development and test environments separate from ArcGIS Server sites can be configured in an
production to maintain reliable operations.
active-passive failover configuration to
Most organizations use multiple support high-availability business needs.
environments to update and maintain their The primary active GIS Server site supports
production operations. Separate production operations, while the secondary
environments are maintained for failover site operates in standby mode
development, staging, and production. configured to take over the production
service operations in the event of primary
• Development – This is a sandbox server site failure.
ArcGIS Server site for testing new
applications and services. Once The active-passive configuration requires
changes are validated on the two separate identical GIS Server site
development site, they are applied to installs, each site with its own ConfigStore
the staging site. and SvrDirectories. A third-party load
• Staging – This ArcGIS Server site is a balancer solution can be configured to
clone of the production site. This site is monitor the primary site and failover to the
used for final performance and secondary site in the event of a primary site
functional testing before moving to the failure. Custom business processes must be
production site. established to maintain the multiple site
configuration. Guidelines provided for

development, staging, and production Each active-passive configuration requires
operations can be used to publish services two separate identical GIS Server site
and maintain data in sync at both site installs, each site with its own ConfigStore
locations. and SvrDirectories. A third-party load
balancer solution can be configured to
ArcGIS Server active-passive failover monitor the primary site and failover to the
configurations within a single data center secondary site in the event of a primary site
location can be supported with a single failure. Custom business procedures must
primary server license – failover secondary be established for building and maintaining
server does not require a separate ArcGIS the multiple site configurations. Guidelines
Server license as long as it operates in a provided for single-machine high-availability
standby mode when the primary server is (active-passive) deployment can be used to
operational. publish services and maintain data at both
multiple site locations.
Data Center COOP/DR Failover

configurations Additional multiple site deployment
considerations
Multiple ArcGIS Server site configurations

are not the optimum solution for most
organizations. There are potential benefits –
multiple site deployment configurations may
adapt well with other Web Service business
processes, particularly in large virtualized
data center operations or cloud
deployments. Multiple site configurations
Figure 9.24 Most critical operations have extend support for operations across
requirements for failover to a separate data center multiple data center locations. For small
location.
organizations, an active-passive failover
solution may reduce licensing costs.
ArcGIS Server sites can be configured in an
active-passive failover configuration to
There is no single best solution for all
support continuity of operation plans
organizations. ArcGIS provides an open
(COOP) and disaster recovery (DR)
platform that supports the simplicity
business needs. The primary active GIS
requirements of most customers, along with
Server site supports production operations,
the flexibility to expand and adapt to meet
while the secondary failover site operates in
the needs of the most advanced
standby mode configure to take over the
organizations. System architecture design
production service operations in the event of
planning can help identify what solution
primary server site failure. Primary site
pattern works best for your organization.
operations can be distributed (shared)
Understanding your business requirements
between the primary and secondary data
and the technology available to satisfy your
center, depending on your business needs.
business needs is critical in building a GIS
that works best in your environment.

Optional high capacity single-machine database or file share. A common service
GIS Server Site active-active output file share can be used to support
configuration tier distributed service requires across the
multiple platform sites.
Single-cluster configurations provide optimum

ArcGIS Server site scalability
ArcGIS Server single cluster sites is the

default configuration since the ArcGIS 10.4
release.
Figure 9.25 Large high capacity data centers may
prefer to build and deploy multiple active-active server The ArcGIS 10.3.1 release provides an
sites. option to remove the cluster aware load
balancing between GIS server machines. A
Single machine ArcGIS Server sites can be single ArcGIS Server site cluster
configured in an active-active high capacity deployment option is preferred over
configuration. Each GIS Server machine deploying multiple active-active server sites.
includes its own ConfigStore and There are several administrative
SvrDirectories. GIS Server machines in this advantages in deploying server machines in
configuration are not cluster aware. Limiting a single site cluster, taking advantage of a
use of built in cluster aware GIS Server site common config store and sharing common
automation will result in a more scalable service configurations.
architecture. ArcGIS 10.3.1 release
provides an alternative ArcGIS Server site Removing GIS Server cluster aware load
high scalability configuration option with balancing between GIS Server machines
cluster aware load balancing turned off. reduces the communication overhead of a
standard ArcGIS Server cluster deployment,
The active-active configuration requires providing scalability comparable with a
multiple GIS Server platform installs. A multi-site server deployment with the
third-party load balancer must be used to advantages of a shared configuration store
assign traffic to the appropriate GIS Server and common data source.
sites and balance processing loads. Custom
business processes must be established to To remove ArcGIS Server cluster aware
maintain the GIS Server multiple site load balancing, the following criteria must
configuration. Guidelines provided for be met:
single-machine high-availability (active-
active) deployment can be used to publish • All GIS servers in the site must
services and maintain data across the participate in a single cluster. Multiple
multiple GIS Server machine farm. clusters cannot exist.
• An external load balancer or ArcGIS
Virtual server clones can be used to build Web Adaptor must be configured to
and automate platform deployment. Service forward requests to the GIS servers in
definition files can be used to deploy and the site. If no external gateway exists,
update services to the production platforms. requests will only be handled by the GIS
Data sources can be replicated to each GIS server designated in the request.
Server platform or maintained in a common

ArcGIS Enterprise server roles
Figure 9.26 ArcGIS Server roles.
Figure 9.26 shows an overview of the • Configured multi-machine site for

ArcGIS 10.5 server roles. ArcGIS Enterprise dynamic image services.
server roles expand support for Portal • Configure multi-machine site for raster
collaboration with geoprocessing and real- analytics.
time services. The ArcGIS Enterprise server
role architecture promotes optimized ArcGIS GeoAnalytics Server
workflow separation configurations.
• Configure multi-machine site for
ArcGIS Server software supports the GeoAnalytics Server.
following installed server configurations
ArcGIS GeoEvent Server
ArcGIS Server core processing capabilities
• Configure single-machine sites for your
• Configured in as many multi-machine particular deployment.
sites that make sense for your particular • At 10.5 and prior: strong
deployment following workflow recommendation to use single-machine
separation recommendations. sites for optimum scalability.
• For example, provide separate sites for
different sets of map services, separate CPT Calculator ArcGIS Enterprise server
sites for heavy-weight geoprocessing, roles.
separate sites for CPU intensive routing
services, ... CPT Design ArcGIS Enterprise server
roles
ArcGIS Image Server

Concluding Remarks
There are several factors that should be considered when establishing your enterprise data
center architecture. Many of these factors are determined based on business needs and
standard IT operating procedures.
The primary focus for Esri system and replication services), system backup,
architecture design services is to identify and security. System migration will normally
hardware and infrastructure resources that include continued support for legacy
satisfy user productivity needs during peak operations while introducing new
GIS system loads. This effort focuses on the technology, often on separate hardware
primary production hardware and available environments.
network infrastructure bandwidth required to
support GIS operations. An Enterprise GIS design includes
business, application, data, and technical
Other factors contribute to the final system architecture requirements. The Capacity
configuration. These factors include Planning Tool provides a framework that
provisions for system maintenance, models enterprise GIS performance and
updates, configuration control, software scalability, integrating the full range of
licensing, and security. System Enterprise system design requirements into
requirements often include hardware a solution that represents your GIS
provisions for application development, production needs.
system test, production staging, background
processing (i.e. map cache maintenance CPT Capacity Planning videos

Performance Management
Fall 2017 Performance Management 41st Edition
Esri has implemented distributed GIS solutions since the late 1980s. For many years,
distributed processing environments were not well understood, and customers relied on the
experience of technical experts to identify hardware requirements to support their
implementation needs. Each technical expert had a different perspective on what hardware
infrastructure might be required for a successful implementation, and recommendations were
not consistent. Many hardware decisions were made based on the size of the project budget,
rather than a clear understanding of user requirements and the appropriate hardware
technology. Many GIS implementation projects would fail due to poor system design and lack of
performance management.
Esri started developing simple system System performance factors

performance models in the early 1990s to
document our understanding about
distributed processing systems. These
system performance models have been
used by Esri system design consultants to
support distributed computing hardware
solutions since 1992. These same
performance models have also been used
to identify potential performance problems
with existing computing environments.
Figure 10.1 Several key system performance factors
The Capacity Planning Tool was introduced work together to provide required user workflow
in 2008 incorporating the best of the productivity. A properly balanced resource investment
traditional client/server and web services will provide the optimum user performance.
sizing models providing an adaptive sizing
methodology to support future enterprise Figure 10.1 identifies some key components
GIS operations. The new capacity planning that contribute to overall system
methodology is much easier to use and performance. Software technology selection
provides metrics to manage performance and application design drives the
compliance during development, initial processing loads and network traffic
implementation, and system delivery. requirements. Hardware and architecture
selection establishing processing
This chapter introduces how these design capabilities and how the processing loads
models can be used for performance are distributed. Network connectivity
management. establishes infrastructure capacity for
handling the required traffic loads.
Warning: Weakest system component

determines overall system performance
(performance chain).
Best practice: Balanced system design

provides optimum user performance at
lowest system cost.
Performance Management 10.1

Software technology factors • Data server including processor core
performance, platform capacity, physical
Software design efficiency and level of memory, and network connection.
analysis establishes complexity of the • Network communications including
bandwidth, traffic, latency, and application
application functions. Data source structure
communication chatter.
and the size and composition of the data
contributes to the complexity of the
The system design solution must provide
information the application must work with.
sufficient platform and network capacity to
process software loads within peak user
Application:
performance needs.
• Core software and client application
Best practice: CPT Standard Workflows
efficiency.
provide proper processing load profile.
• Display complexity includes layers per
display, features per display extent,
functions used to complete the display, and How is performance managed?
display design for each map scale.
• Display traffic System architecture design provides a
• User workflow activity including user
framework for identifying a balanced system
productivity, implementation of heavy
workflow tasks, and workflow efficiency
design and establishing reasonable
(mouse clicks to final display, software processing performance budgets.
communication chatter) Performance expectations are established
based on selected software processing
Data source: complexity and vendor published hardware
processing capacity. System design
• Data source technology including DBMS
performance expectations can be
(data types, indexing, tuning, scalability), file represented by established software
source (File format, structure, indexing, processing performance targets. These
scalability), imagery (Image format, file size, performance targets can be translated into
indexing, pre-processing, on-the-fly specific software performance milestones
processing), or cached data source. which can be validated during system
• Geodatabase design including table deployment. Software processing
structure, dependencies, and relationship complexity and/or hardware processing
classes. capacity can be reviewed and adjusted as
• Data connection including SDE (direct necessary at each deployment milestone to
connect, applications server connect) or file ensure system is delivered within the
source (internal disk, direct attached,
network attached).
established performance budget.
Hardware technology factors Our understanding of GIS processing

complexity and how this workload is
supported by vendor platform technology is
Hardware design and performance
based on more than 20 years of experience.
characteristics determine how fast the
A balanced software and hardware
servers can do work and the volume of work
investment, with capacity based on
they can handle at one time.
projected peak user workflow loads, can
reduce cost and ensure system deployment
• Workstation/application server/GIS server success.
including processor core performance,
platform capacity (servers), physical
memory, network connection, graphics
processing unit.

processing than medium complexity
workflows.
• Additional complexity selections (2x
medium, 3x medium, 4x medium, …10x
medium) are available for establishing much
heavier performance targets.
Workflow complexity guidelines:
• Light complexity is the minimum loads

expected based on software technology
selection.
• Medium complexity would support up to 80
percent of selected software technology
Figure 10.2 Performance management involves
deployments.
building a design solution based on appropriate
• Heavy complexity represents user workflows
workflow performance targets and managing
compliance throughout design and implementation to
with more complex data models (more
deliver within those targets.
layers, more features per layer, and more
complex analysis).
Most project managers clearly understand

• 2x medium, 3x medium, 4x medium, …10x
the importance and value of a project medium represent much more complex
schedule in managing deployment risk workflow loads that are possible with
associated with cost and schedule. The expanding technology and emerging display
same basic project management principals details.
can be applied to managing system
performance risk. Figure 10.2 shows some Faster hardware processing allows more
basic concepts that can be used in complex analysis to be included in the user
managing performance. workflows. These heavier complexity
workflows (2x, 3x, 4x, …10x medium) may
System architecture design framework: not handle a large number of concurrent
users, but with today's technology they can
• CPT provides balanced standard and deliver map display results in a reasonably
custom workflow load profiles. response time (less than 5-sec).
• Workflow complexity assessment is used to
assign reasonable software processing Best practice: Performance expectations are
performance budgets. established based on selected software
processing complexity and vendor published
Workflow complexity assessment: hardware processing speed (per core
performance).
• Light complexity represents simple user
displays with minimum functional analysis
(light processing loads).
• Medium complexity represents standard
workflow performance targets that satisfy
most workflows that apply best practice
design standards. Medium complexity is
roughly twice light complexity processing
loads.
• Heavy complexity represents workflows that
include more complex map displays or data
models that generate 50 percent more

Platform throughput and The processor core is the hardware that
executes the computer program
service time instructions.
• Number of processor core identifies how

many instances can be serviced at the same
time.
• Service time is a measure of the average
work transaction processing time.
Work transaction service time is a key term

used to measure software performance.
• The software program provides a set of

instructions that must be executed by the
computer to complete a work transaction.
• The processor core executes the
Figure 10.10 provides a chart showing the instructions defined in the computer program
relationship between utilization and throughput; a to complete the work transaction.
simple relationship that can be used to identify
platform capacity.
Transactions with more instructions
represent more work for the computer, while
The most important system performance transactions with fewer instructions
terms define the average work transaction represent less work for the computer.
(display), work throughput, system capacity,
and system utilization. Figure 10.10 The complexity of the computer program
provides a chart showing the relationship workflow can be defined by the amount of
between utilization and throughput; a simple work (or processing time) required to
relationship that can be used to identify complete an average work transaction.
platform capacity.
• Service time on the CPT Workflow tab is
Capacity (DPM) = Throughput presented relative to a platform performance
(DPM)/Utilization baseline.
• Faster platform processor cores execute
Best practice: If you know the current program instructions in less time than slower
throughput (users working on the system) processor cores.
and you measure the system utilization • Service time can be computed using a
(average computer CPU utilization), then you simple formula based on number of
can know the capacity of the server. processor cores and platform capacity.
The relationship between throughput, Service time (sec) = 60 sec x

capacity, and utilization are true based on #core/Capacity (DPM)
how these terms are defined.
Service time can be computed based on
• Throughput is the number of work measured throughput and utilization.
transactions being processed per unit time.
• Capacity is the maximum throughput that
can be supported by a specific hardware
configuration.
• Utilization is the ratio of the current
throughput to the system capacity
(expressed as percentage of capacity).

Figure 10.13 provides a chart showing the
relationship between utilization and
response time.
You can calculate display service time if you

know the platform throughput and
Figure 10.11 Service time calculations for peak loads
corresponding utilization, calculated at any
generated at each web service instance configuration.
throughput level. Calculating user display
Service time calculations response time for shared system loads is a
little bit more difficult.
Figure 10.11 shows service time results for
five different throughput loads. Calculating user response time:
• Number of deployed service instances • Only one user transaction can be serviced at
determine peak loads. a time on each processor core.
• Throughput and utilization are measured for • If many user transaction requests arrive at
each of the five separate test configurations. the same time, some of the transactions
• Capacity of 714 DPM was calculated from must wait in line while the others are
each test load. processed first.
• Service time of 0.34 sec was calculated from • Waiting in line for processing contributes to
each test load. system processing delays.
• User display response time must include
Best practice: You can calculate capacity time for all the system component
processing times and system delays, since
from throughput and utilization
the display is not complete until the final
measurements at any system load.
processing is done.
Note: Real operational environments can

Any system time where a transaction
provide a very good measure of capacity.
request must wait in line for processing is
called queue time.
Once you know the platform capacity, you
can compute the platform service time. Response time is the sum of the total
service times (processing times) and queue
times (wait times) as the transaction request
Platform performance and travels across system components to the
response time server and returns to deliver the final user
display.
Response time (sec) = Service time (sec) +

Queue time (sec)
Warning: Queue time increases to infinity as

any processing component of the system
approaches full capacity.
Response time is importance, since it

directly contributes to user productivity.
Figure 10.13 Display response time increases with Productivity = 60 sec/(response time + think
increased platform loads. time)

Warning: As queue time increases response • Network queue time = network congestion
time will increase and productivity will delays
decrease. • Network latency delay time = measured
latency (round trip travel time) x chatter
(round trips)
How to size the network
Best practice: CPT includes network as
additional system component when
computing system performance.
Warning: Network performance can be the
most critical design constraint for many
distributed system design solutions.
What is a valid user workflow?
Figure 10.15 Display response time increases with

increased network loads.
Figure 10.15 provides a chart showing the Figure 10.7 A valid workflow provide sufficient time for
relationship between network utilization and user to review the display and enter the following
response time. Performance models used to display request.
support network communications follow the
same type of terms and relationships Figure 10.7 shows a valid workflow. All user
identified for server platforms. workflow performance terms work together
during each display transaction to satisfy
Some of the same performance terms are business performance requirements.
referenced by different names.
Workflow specifications:
• Network transaction = display
• Network throughput = traffic • User productivity = 10 DPM/client (user
• Network capacity = bandwidth workflow performance needs)
• Network utilization = utilization • Display cycle time = 6 sec (60 seconds in a
minute divided by 10)
The network connection (switch port, router
port, network interface card, hardware bus For a given display executed on a given
adapter, etc.) is the hardware that platform:
processes the network traffic.
• Display service time is a constant value.
• Most local networks are identified as single • In a shared server environment, queue time
path systems. increases with increasing user loads
• Multiple NIC cards or multiple network paths (increasing server utilization).
can improve throughput utilization. • As queue time increases, display response
time increases.
Additional performance terms: • For a fixed user productivity (10 displays per
minute), computed user think time will
decrease with increasing display response
• Network service time = network transport time.
time

Computed user think time is greater than CPT Design ADJUST process:
minimum think time for valid workflow.
• Iterative calculation that reduces user
Warning: At some point, computed user productivity for all invalid workflows and then
think time will be less than minimum think re-computes the system solution.
time (invalid user workflow). • If adjusted productivity provides minimum
think time less than computed think time, the
next iteration will increase productivity
User productivity adjustment
slightly and re-compute the system solution.
• Iterations continue until the most critical
adjusted computed think time = minimum
think time.
Best practice: Enable iterative calculations in

Excel Options > Formula.
o Maximum Iterations: 500

o Maximum Change: 0.001
Warning: Excel will provide a Circular

Reference Warning if the Enable iterative
Figure 10.8 The CPT Design identifies an invalid calculations is not selected. Iterative
workflow when computed think time is less than calculations are required for many of the
minimum think time. The CPT Adjust function reduces CPT sizing calculations.
user productivity value until computed think time =
minimum think time. Workflow is valid once computed CPT workflow productivity adjustment
time is equal to or greater than minimum think time.
Best practice: System design should be
During peak system loads, queue time can upgraded to satisfy user productivity needs.
increase to a point where computed think
time is less than minimum think time as Geoprocessing services (batch
shown in Figure 10.8. The user productivity workflows)
must be adjusted (reduced) to represent a
valid user productivity.
A batch process is a workflow that does not
require user interaction. User inputs are
CPT identifies an invalid workflow by
provided before the process is executed.
changing the workflow productivity.
The process then runs without user input
until the job is done. Figure 10.9 shows a
• Workflow productivity must be reduced to diagram representing a batch process.
identify a valid workflow.
• CPT includes a RESET ADJUST function
that will automatically reduce workflow
productivity to the proper reduced value.
CPT Design ADJUST function:
• Valid system solution is reached when

computed user think time is equal to or
greater than minimum think time for all
workflows. Figure 10.9 Batch process loads are sequential in
• Valid solution is identified on the CPT nature and productivity depends on computed
display once valid workflow is established. response time.

Most heavy GIS functions can be modeled • Heavy geospatial analysis
as a batch process. GIS heavy batch • Heavy network analysis
processes, when deployed on Server, are
often called geoprocessing services. Best practice: Any heavy system-level
geoprocessing function that may be
Geoprocessing functions can be deployed requested by more than one user at a time
as a network service configured to handle should be separated from the user
multiple user service work requests. application workflows and executed as
separate network batch process work
• Geoprocessing function runs as a sequential request services.
batch process.
• Each concurrent geoprocessing instance Warning: CPT Design productivity adjust
consumes a single platform core. function must be used to computer system
loads and batch process productivity. Each
Advantages of configuring geoprocessing concurrent batch process are identified in
functions as a network service: the CPT Design as a user (column C) or
client (column D) instance.
• Service work request is sent to a processing CPT representation of batch processing loads
queue to await execution.
• Specific number of server cores can be Best practice: Workflow selection should
allocated to execute the service. have same load profile (client, web, GIS
• User can do other work while waiting for the server, SDE, DBMS) as the batch process
work request to be serviced. you wish to model. Total processing time is
not important for modeling load profile.
Batch process loads are modeled as a
workflow with zero (0) think time (no user The batch process productivity must be
input between display transactions). computed to identify a valid workflow.
Productivity will depend on the server loads
• Batch productivity is calculated based on and available system resources. A single
computed response time (60 batch process can take advantage of only
seconds/response time = batch DPM). one processor core.
• Batch process queue time is limited to
service contention (no random arrival queue
time). Best practice: Recommended design
• Displays are requested sequentially practice - any heavy function (runs more
following each refresh. than 30 seconds) that might be requested by
• Batch processes deployed on a single several users at a time should be configured
platform with local data source tend to as a batch process (network services).
consume a single processor core. Processing queue must be established for
• CPT Design tab will distribute loads across user work request input. Each batch instance
available cores resources based on batch (network service) will process requests
workflow profile (limiting system component sequentially based on available processor
will determine peak batch productivity). resources. User can be notified once their
work request is services.
Batch processing examples:
Platform queue time
• Map caching
• Enterprise Geodatabase reconcile and post
Computing response time is a common
• Geodatabase replication
problem for many business applications. To
• Heavy map printing jobs
• Heavy routing analysis
get it right, you have to understand queue
• Heavy imagery processing time. The theory of queues or waiting in line

has its origin in the work of A. K. Erlang, Queue time model
starting in 1909.
The single-core platform queue time
increases with increasing service time and
platform utilization.
Queue time (single-core) = service time

(sec) x utilization/(1 - utilization).
Queue time is zero (0) when utilization is

zero (0) and increases to infinity as
utilization approaches 100 percent.
In the multi-core platform case, it is

important to include the probability of a
processor core being available to service
Figure 10.14 Transaction request queue time will vary
with platform utilization and number of platform core.
the request on arrival(not busy).
• The more processor cores in the server, the

Figure 10.14 shows a formula for queue
more likely one of these cores will be
time and a graph showing the relationship available for processing when the service
between queue time and platform utilization. transaction arrives.
The number of platform processor core • The equation simplifies to the simple single-
determines the sensitivity of queue time to core formula when the number of processor
platform utilization. cores = 1.
The simplest queuing models work for large Multi-core availability = 1/{1 + utilization x
populations of random arrival transactions, (cores - 1)}
which should certainly be the case when
modeling computer computations Queue time = Multi-core availability x
(thousands of random computer program Queue time (single-core)
instructions being executed within a
relatively small period of time—e.g., The derived queue time formula provided above
seconds). has been compared against several benchmark
test results, and the computed response time
The queue time calculations used in the was reasonably close to the measure test
Capacity Planning Tool is a simplified model results (shows conservative response times—
developed from Operations Research slightly higher than measured values).
Queuing theory.
It is important to recognize that the accuracy
• The second half of the model (single core of the queue time calculation impacts only
section) is quite straight forward, and there the expected user response time, and does
is general agreement that this simple model not reduce the accuracy of the platform
would identify wait times in the case of a capacity calculations provided by the earlier
single service provider (single core platform
simple relationships.
or single network connection).
• The multi-core case is a little more
complicated, and unfortunately is the more • For many years, Esri capacity planning
common capacity planning calculations we models did not include estimates for user
need to deal with in multi-core server response time.
platform configurations.

• Workflow response time is important, since it service time and queue time contributions for
directly impacts user productivity and components across the distributed system
workflow validity. environment.
• If display response times are too slow, the
peak throughput estimates would not be Figure 10.16 shows the information
achieved and the capacity estimates would provided by the CPT Workflow Performance
not be conservative. Summary. Workflow service times and
queue times are shown in a stacked bar
Best practice: Including user response time chart. Response time, shown at the height
in the capacity planning models provides of the stack, is the total time required to
more accurate and conservative platform complete the work transaction.
specifications, and gives customers with a
better understanding of user performance
The Workflow Performance Summary chart
and productivity.
shows the performance of 10 separate
benchmark tests.
Queue time derivatives
• Test were performed on 2-core servers.
Display response time based on percent • Number of concurrent batch processes was
utilization increased with each test run.
• First two tests (1 and 2 batch processes)
Multi-core servers provide better response response time was about the same.
times than single-core servers during heavy • Response time increased linearly for tests
loads. with more than 2 batch processes.
• Eight 1-core servers at 80 percent utilization Response time includes all of the
provide over 4-second response time. processing times and queue times
• Four 2-core servers at 80 percent utilization experienced in completing an average work
provide 3.2-second response time. transaction.
• Two 4-core servers at 80 percent utilization
provide 2.3-second response time. • Platform service and queue times
• One 8-core server at 80 percent utilization • Network transport and queue times
provide less than 1.6-second response time. • Latency travel time delays
• Client service time
Warning: More cores per server improves
response times only when display service
times are the same for all configurations.
CPT Design multi-core platform performance
demonstration
Server deployment transaction
What is system performance? throughput capacity constraints
Several technology factors impact
performance and scalability of deployed
server systems. Selecting the optimum
configuration strategy will help ensure peak
system throughput and optimum return on
investment. The following technology
factors are important in developing an
optimum ArcGIS deployment solution.
ArcGIS Server Site processing overhead

Figure 10.16 System performance must consider

Virtual Server consolidation
Figure 10.18 ArcGIS Server deployed in a physical

server architecture.
Figure 10.17 ArcGIS Server Site communication

overhead will vary with transaction throughput and
Figure 10.18 shows a typical Enterprise GIS
number of GIS Server machines. production environment supported by a
physical server architecture.
ArcGIS Server provides a cluster aware
capability that supports deployment of For many years, data centers were
multiple clustered server machines within a supported by physical server configurations.
single ArcGIS Server site. Each GIS Server With physical server deployment
machine communicates with a common
configuration store and with each machine • Many servers were required to support
in the site for load balancing and service Enterprise operations.
transaction assignment. ArcGIS Server • Many servers were performing well below
their optimum capacity.
cluster aware processing overhead varies
• High number of servers contributed to data
based on the transaction throughput rate center high power consumption.
and the number of machines in the ArcGIS
Server site.
Figure 10.17 shows the peak transaction

throughput based on number of machines in
the ArcGIS Server site. Five services with
different levels of complexity are shown in
the graphic. The results demonstrate that
Figure 10.19 ArcGIS Server deployed in a physical
very light services with very high transaction server architecture.
rates do not scale out well due to the
multiple machine site communication
Figure 10.19 shows a typical Enterprise GIS
overhead.
production environment supported by a
virtual server architecture. Virtualization
CPT Design demonstration of ArcGIS Server
reduces the total number of data center
Site scalability
physical servers.
ArcGIS 10.3.1 for Server provides an option
Virtual server machines are deployed on
to remove internal load balancing for GIS
host server platforms.
Server site machines for a more scalable
configuration. This siloed deployment option
• Multiple virtual machines can be supported
(internal site communications are disabled)
by a single host server configuration.
is limited to single cluster sites and provides • Host platforms can run at optimum capacity
linear scalability. levels (50 percent to 80 percent utilization).
• Virtual Server architecture can be deployed
CPT Design demonstration of ArcGIS Server to optimize host platform processing loads.
single-cluster site scalability

Virtualization: Host server processing approaches 100 percent, the VM utilization
loads will be limited based on available host
resources.
CPT Design demonstration of ArcGIS Server

Virtual Machine (VM) performance.
Available Virtual Server machine utilization

and throughput is limited by hypervisor
processing overhead when virtual servers
must compete with available host platform
processing resources.
Figure 10.20 Virtual Server host machine hypervisor
processing overhead. Best practice: Provide host platform with 50
percent more processing capacity that
required by the virtual servers.
Figure 10.20 shows the hypervisor
processing loads supported on the host
platform and the impact on virtual server Esri/VMware joint benchmark testing
utilization. reports.
Virtual Server machines (VM) are deployed • October 2011 Esri ArcGIS Server 10 for
VMware Infrastructure Deployment and
on a host platform, with access to
Technical Considerations Guide includes
processing resources controlled by a performance testing of ArcGIS Server 10
hypervisor. The hypervisor assigns VM with VMware ESXi 3.5u4.
virtual core to host platform hardware CPU • July 2013 Esri ArcGIS 10.1+ for Server on
resources, allocating available processing VMware vSphere Deployment and Technical
resources between the deployed VMs. Considerations Guide includes performance
testing of ArcGIS 10.1 for Server with
Hypervisor processing loads are supported VMware vSphere 5.1.
directly by the host platform and can be
serviced by available host CPU resources Test results show significant virtual server
separate from the CPU resources assigned performance improvements with the more
to Virtual Server machines (if extra CPU recent VMware vSphere technology. The
resources are available). When host October 2011 testing showed slightly more
platform CPU resources are limited, the than 10 percent virtual server processing
hypervisor must compete with the VM core overhead per core, while the July 2013
for access to available host platform testing showed limited performance
resources. degradation between physical and virtual
server deployment configurations when the
Test results show hypervisor loads may virtual host platform performs at levels less
account for up to 50 percent of the total than 90 percent utilization.
virtual server processing loads. Virtual core
for each VM must be assigned to available Note: July 2013 testing showed virtual server
host platform physical core for processing. hypervisor overhead of 30 percent running on
Optimum VM throughput is achieved when the host platform (50 percent of the VM loads).
sufficient host resources are available to
support all VM processing requests along Performance Validation
with the hypervisor processing load without
having to compete for processing
resources. As host platform utilization

Planning provides the first opportunity for Figure 10.21 shows three key opportunities
building successful GIS operations. Getting for measuring performance compliance.
started right, understanding your business When possible it is important to take
needs, understanding how to translate advantage of opportunities throughout
business needs to network and platform system development and deployment where
loads, and establishing a system design that you can measure progress toward meeting
will satisfy peak user workflow requirements your performance goals. The CPT Test tab
is the first step on your road to success. includes four tools you can use to translate
live performance measurements to workflow
Planning is an important first step – but it is service times – the workflow performance
not enough to ensure success. If you want targets used to define your initial system
to deliver a project within the initial planning design.
budget, you need to identify opportunities
along the way to measure progress toward Map display render times
your implementation goal. Compliance with
performance goals should be tracked In Chapter 3 we shared the important
throughout initial development, integration, factors that impact software performance.
and deployment - integrate performance For Web mapping workflows, map
validation measurements along the way. complexity is the primary performance
Project success is achieved by tracking step driver. Heavy map displays (lots of dynamic
by step progress toward your map layers and features included in each
implementation goal, making appropriate map extent) contribute to heavy server
adjustments along the way to deliver the processing loads and network traffic. Simple
final system within the planned project maps generate lighter server loads and
budget. The goal is to identify problems and provided users with much quicker display
provide solutions along the way - the earlier performance. The first opportunity for
you identify a problem the easier it will be to building high performance map services is
fix. System performance can be managed when you are authoring the map display.
like any other project task. We showed how
to address software performance in Chapter There are two map rendering tools available
3, network performance in Chapter 5, and on the CPT Test tab that use measured
platform performance in Chapter 7. If you map rendering time to estimate equivalent
don’t measure your progress as these workflow service times. One tool is available
pieces come together, you will miss the for translating ArcGIS Desktop map
opportunity to identify and make the rendering times (MXD) and the other tool is
appropriate adjustments needed to ensure for translating ArcGIS Server map service
success. rendering times (MSD). With both tools,
measured map rendering time is translated
to workflow services times that can be used
by the CPT Calculator and Design tabs for
generating your platform solution. The idea
is to validate that your map service will
perform within your planned system budget
by comparing the workflow service times
generated from your measured rendering
times with your initial workflow performance
targets. If the service times exceed your
planned budget, you should either adjust
Figure 10.21 Manage performance by sharing display the map display complexity to perform
complexity targets and measuring compliance within the initial planning budget or increase
throughout system deployment.

your system performance budget. The best Best practice: Performance metrics can be
time to make the map display complexity collected from benchmark test or live
adjustment is during the map authoring operations.
process. Impacts on the project budget can
Warning: Make sure all measurements are
be evaluated and proper adjustments made
collected for the same loads at the same
to ensure delivery success.
time.
Map publishing preview render times

System monitor concurrent users and
platform utilization
Measured MSD render time
Measured peak concurrent users and platform
MSD render time can be measured when utilization translator
publishing your map service using the
service editor preview tool.
If you don’t have measured throughput,
concurrent users working on the system can
Warning: Make sure to measure a map
be used to estimate throughput loads. This
location that represents the average map
is a valuable tool for using real business
complexity or higher within your service area
activity to validate system capacity
extent.
(business units identify peak user loads and
IT staff identify server utilization observed
MXDPerfStat render times during these loads). The Test tab can be
used to input throughput (peak concurrent
Measured MXD render time users), the platform configuration (server
platform selection), and the measured
MXD render time can be measured using platform utilization and excel will translate
the [MXDperfstat] ArcScript performance these inputs to equivalent workflow service
measurement tool. times.
Warning: Make sure to measure a map Best practice: Analysis assumes peak users
location that represents the average map are working at web power user productivity
complexity or higher within your service area (6 DPM) over a reasonable measurement
extent. period (10 minutes).
System test measured throughput and Warning: Make sure all measurements are
platform utilization collected for the same loads at the same
time.
Measured throughput and platform utilization Move Test tab derived workflow service times to
project workflows.
If you know your platform configuration,
your measured peak workflow throughput, The CPT Workflow tab is where the results
and the associated platform utilization the of your performance validation efforts come
CPT can calculate the workflow service together. You can bring all your test results
times. The Test tab translation tools can be together, along with the original workflow
used to input throughput (transaction per service times, to validate that you are
hour), the platform configuration (server building a system that will perform and scale
platform selection), and the measured within your established project performance
platform utilization and excel will translate budget.
these inputs to equivalent workflow service
times. Best practice: Performance management,
including performance validation throughout

development and system delivery, is the key tools can be used to translate real
to implementation success. It is important performance measurements to equivalent
that you identify the right technology and workflow service times for performance
establish reasonable performance goals validation.
during your initial system design planning. It
is even more important that you monitor
progress in meeting these goals throughout
final system development and delivery.
Capacity Planning
The models supporting Esri capacity
planning today are based on the
performance fundamentals introduced in
this section. Platform capacity is determined
by the software processing time (platform
service time) and the number of platform
core, and is expressed in terms of peak
displays per minute. Platform capacity
(DPM) can be translated to supported
concurrent users by dividing by the user
productivity (DPM/client).
The performance fundamentals discussed

in this chapter are basic concepts that apply
to any computer environment, and an
understanding of these fundamentals can
establish a solid foundation for
understanding system performance and
scalability. Software and hardware
technology will continue to change, and the
terms and relationships identified in this
section can be used to normalize these
changes and help us understand what is
required to support our system performance
needs.
The next chapter will provide an overview of

the Capacity Planning tools introduced
throughout the previous chapters. The CPT
videos at the end of this chapter focus on
system performance validation – showing
how the fundamental performance terms
and relationships are used by the CPT to
connect user requirements with system
hardware loads, and how these loads are
used to identify appropriate hardware
requirements. Performance validation
during system design and deployment is
also a key topic, sharing how the CPT Test

City of Rome
Fall 2017 City of Rome 41st Edition
This chapter shares a process you can use to complete your own system design. This process
brings together what has been discussed in the earlier chapters and demonstrates the value of
the system architecture analysis in making informed design decisions.
System design provides a methodology for establishing hardware and network requirements
that support the performance and communication needs of GIS application users. Hardware
requirements should be established based on identified business needs. A fundamental
understanding of user workflow requirements (business architecture) and the supporting GIS
technology is required before one can identify the appropriate hardware and network
requirements for supporting effective enterprise GIS operations.
City of Rome is the name of the case study Figure 11.1 shows a collection of photos
provided to demonstrate the planning representing City of Rome. The fictional City
process presented in a book by Roger of Rome represents a typical organization,
Tomlinson called Thinking about GIS: just right as a case study to demonstrate
Geographic Information System Planning how you can use the capacity planning tool
for Managers. Both his book’s chapter 9 and in your system design process.
this chapter show standard templates that
can be used for most enterprise design Phase 1 of the case study will implement an
studies. The City of Rome case study is ArcGIS Enterprise solution to enhance the
updated each year to represent current existing business operations.
technology trends and ArcGIS migration
strategies. In this chapter, we will use the • Model the existing City of Rome
Capacity Planning Tool as a framework to operations
model user requirements and the system • ArcGIS Enterprise IOC architecture
design for two planned phases of expansion • Vector tile basemap architecture
and growth for the City of Rome. • Citrix host hardware platform upgrade
• ArcGIS Enterprise hardware upgrade
baseline
City of Rome case study
Best practice: No additional software
licensing is required to upgrade existing
ArcGIS operations to include ArcGIS
Enterprise IOC capabilities.
Phase 2 will implement a Water Utility

solution for City of Rome
• Water solutions user needs analysis

• Water solutions system design
Figure 11.1 The City of Rome is a typical municipal
community that is used to represent how you might Best practice: CPT will be used to
use the system design methodology to make your complete the system design analysis.
infrastructure upgrade decisions.
City of Rome 11.1

Pre-design efforts
Figure 11.2 Business needs establish the foundation for any enterprise GIS design. The enterprise vision, existing
business architecture, and user requirements must be understood to select the best GIS solution.
Figure 11.2 shows the efforts completed in • Operational constraints and priorities
preparation for the system design. Business • Funding constraints
needs must be understood before you are
ready to complete the system design. Workflow loads analysis. User workflow
loads analysis reviews the business
Enterprise vision. GIS software processes to identify where and what is
deployment patterns are optimized to required to support business needs.
support your business needs:
• user location and connectivity
• Location enablement • user workflow analysis (user needs)
• Data management
• Analysis
• Field mobility
City of Rome existing
• Visualization operations
• Constituent engagement
For this case study, the existing City of
Existing Business Architecture. Business Rome GIS Operations will first be modeled
architecture defines the current state of how to establish a design baseline. Then the
you are meeting your business "existing operations" design will be
requirements. expanded to include a Phase 1 ArcGIS
Enterprise initial operational capability
• Governance and political landscape (IOC). This will be followed by review and
• People and communication strategies deployment of a Phase 2 Water Utilities
• Platform and network environments solution for City of Rome.
City of Rome 11.2

Workflow loads analysis: Existing operations
Figure 11.3 City of Rome existing GIS operations user needs summary.
Figure 11.3 shows the City of Rome existing • Peak loads will be used to identify
workflow loads analysis, providing a hardware and network system
summary of user locations and peak user design specifications.
workflow loads.
Best practice: CPT will be used to
The workflow loads analysis includes all the complete the system design analysis.
GIS workflow patterns identified during the
business needs analysis.
User workflow display complexity:
• A simple spreadsheet layout can Existing operations
show user locations by department
and site location.
• Common workflow technology
patterns can be combined across
site locations to simplify the display
and design analysis.
• The peak usage estimates represent
a snapshot of a peak load profile
that will be used for the system
design. Figure 11.4 Existing business workflow patterns
• Workflows represent the technology
profiles that will be used during peak Figure 11.4 shows the workflow patterns
operations. identified for the City of Rome existing
operations. User workflow patterns are
City of Rome 11.3

normally a product of a user needs • Estimate the average complexity of
assessment and provide a reference for each user workflow.
establishing workflow loads for the system
design. User workflow complexity is an estimate of
the processing loads required for an
The software technology workflow patterns average information product display
are identified during the workflow loads transaction. The workflow complexity
analysis. Follow these recommendations estimate will be used with the CPT to
when performing a user workflow loads generate appropriate system design
analysis: performance specifications.
• Review each user business case to Best practice: A conservative workflow

identify appropriate workflow categories. complexity will ensure an adequate
• Identify the required user workflows and system design solution
solution architecture for each technology
pattern.
Workflow performance targets: Existing operations
Figure 11.5 Workflow performance targets for the existing City of Rome GIS operations.
Figure 11.5 shows the workflow technology performance factors for each
performance targets selected for the City of use workflow.
Rome existing operations. The CPT can be
used to generate appropriate workflow The workflow definition provides a recipe
performance targets for the system design. that identifies the software technology
performance factors used to generate each
Workflow performance targets are user workflow performance target.
generated by the CPT Calculator by
selecting the appropriate software Best practice: The user needs workflow
technology performance factors. The loads analysis summary and the user
software technology profile and complexity workflow performance targets are used
identified during the user needs workflow by the CPT to generate traffic and
loads analysis provide information needed processing loads on the CPT system
to select the appropriate software design solution.
City of Rome 11.4

CPT Workflow Loads Analysis: Existing locations, network bandwidth
operations connections, and central data center
locations. Drawing should match the
CPT project workflows: Existing operations
user location information provided on
the user requirements templates.
• User requirements analysis: The CPT
System design process: Existing Requirements analysis section is
operations configured to represent the site
locations, user workflows, peak loads,
and network bandwidth for the
enterprise design solution.
• Network suitability analysis: CPT Design
completes the network suitability
analysis and identifies any
communication bottlenecks. Network
bandwidth upgrades are identified to
complete the network suitability
analysis.
• Platform architecture selection: The
CPT Design Platform tier is configured
to represent the design solution. Identify
platform tier nicknames, select
Figure 11.6 System design process provides a logical platforms, and identify platform rollover
step-by-step methodology for using the CPT to
complete your system design.
settings.
• Software configuration: The CPT Design
Software Configuration module is used
Figure 11.6 shows the process used to
to assign workflow software to
complete the system design. The initial
supporting platform tier (software install)
effort for the City of Rome design is to
and make workflow data source
establish a baseline planning environment.
selection.
The standard system design process
• Enterprise design solution: Once
introduced in Chapter 1 will be used to
configured, the CPT Design tab
model the existing City of Rome
completes the system design analysis
environment.
and provides the platform solution.
System design process:
• Technical architecture strategy. High-

level network drawing showing user site
City of Rome 11.5

Technical architecture: Existing operations
Figure 11.7 City of Rome existing technical architecture and network connectivity.
Figure 11.7 shows the City of Rome o Site 4 Willsberg—45 Mbps WAN
technical architecture providing a summary connection
of the current user locations and network • Data center—45 Mbps Internet
connectivity. connection
o Public web services will connect
The GIS server machines are supported in through the data center Internet
the central IT data center in City Hall. connection.
• Windows Terminal Server farm for Current network cost (WAN/Internet

remote ArcGIS Desktop users. Service) is provided for reference purposes
• Web server machines for internal and ($7,836/month).
public web services.
• ArcGIS Server machines for mapping Warning: The estimated network pricing
services. used in the CPT is for demonstration
• Data server machines for GIS data purposes only.
sources.
Network pricing model
City Hall data center remote network
connections are as follows:
• Data center—1000 Mbps LAN

connection
• Data center—90 Mbps WAN connection
o Site 2 Operations facility—12
Mbps WAN connection
o Site 3 Freeberg—45 Mbps WAN
connection Figure 11.8 Network pricing model for Wide Area
Network (WAN) and Internet service provider rates.
City of Rome 11.6

Figure 11.8 shows the network service Warning: Network service rates can be a
provider cost model used for demonstration significant performance/cost
purposes in this case study. consideration for an optimum design.
Best practice: Network bandwidth
Network service provider costs should be upgrade recommendations must be
considered when evaluating Enterprise level coordinated with the network
GIS design solutions. administrator and included in the
network infrastructure upgrade budget.
• Network bandwidth has an impact on
remote user productivity with every GIS CPT Design requirements analysis:
design. Existing operations
• Technology change (i.e., higher
resolution clients) increase GIS CPT Design network suitability analysis:
workflow traffic requirements. Existing operations
• Dynamic feature services generate
more traffic than standard mapping
services.
Platform architecture selection: Existing operations
Figure 11.9 GIS data center platform architecture for existing City of Rome GIS operations.
Figure 11.9 shows City of Rome existing availability virtual server platform
platform architecture components. Each architecture.
server machine will be supported by a high-
City of Rome 11.7

The existing platform architecture includes Figure 11.10 shows the CPT Design
the following virtual machines. workflow performance summary for the
existing GIS operations.
• WTS tier: ArcGIS Desktop (ArcMap)
remote clients supported by XenApp Establishing a baseline model of existing
sessions on the Citrix host platform tier. operations validates user needs estimates
• Web tier: A total of four (4) web servers and provides quality assurance for
(two internal and two public) support the supporting system design for future
high-availability Web tier architecture. deployments.
• GIS tier: A total of four (4) ArcGIS
Servers (separate internal and public Workflow performance summary provides
GIS Server sites, each with two the following:
machines to support high-availability
requirements). • Average processing and queue times for
• DBMS tier: A total of four (4) DBMS each user workflow.
servers (primary production and • Average difference in performance
publication server machines each with a between each site location.
dedicated failover server to satisfy high- • Average performance for shared public
availability requirements). services.
Best practice: A virtual server data Important to note that model shows average
center environment provides optimum performance levels.
flexibility for deploying and supporting
GIS operations. • Some display response times can be
lighter.
CPT Design platform configuration: • Other display response times will be
Existing operations slower.
• Queue times will vary depending on
CPT Design software configuration: random load profiles.
Existing operations
Best practice: Workflow performance
Workflow performance summary: Existing summary can be used to identify
operations
expected relative performance between
remote locations and establish user
experience expectations for the
configured design solution.
Figure 11.10 Workflow performance summary for the

existing City of Rome business workflows.
City of Rome 11.8

Enterprise design solution: Existing operations
Figure 11.11 City of Rome existing high-availability virtual server platform solution.
Figure 11.11 shows the final platform design The physical host server environment is
solution. After you finish configuring the supported by six (6) existing 2011 Xeon
CPT Design tab, Excel completes the X5687 8-core platforms.
system design analysis and provides a
diagram of the platform solution.
CPT Design platform solution: Existing operations

Phase 1 ArcGIS Enterprise
IOC
This graphic shows the 4-tier virtual server
platform architecture configuration along The City of Rome initial software migration
with the CPT generated host platform tier will deploy ArcGIS Enterprise components
loads for the existing City of Rome GIS to establish an ArcGIS Enterprise Initial
operations. The CPT Design analysis has Operational Capability (IOC). This
loaded the selected platform solution with deployment will expand existing operations
the peak user loads identified in the design. to include a system of engagement for
existing users. The ArcGIS Enterprise IOC
The following host platform tier support the deployment will leverage existing hardware
data center processing loads. and ArcGIS user licensing. The ArcGIS
Enterprise deployment will enable existing
• Citrix host platform: Three platform users to create content (web maps),
nodes at 34.0% utilization. collaborate and share ideas, and use
• Server host platform: Three platform configurable apps to empower more
nodes at 31.6% utilization. adaptive Web GIS operations.
City of Rome 11.9

Workflow loads analysis: ArcGIS Enterprise IOC
Figure 11.12 City of Rome ArcGIS Enterprise IOC user needs summary.
Figure 11.12 shows a projected user needs Expanded GIS operations

summary following the ArcGIS Enterprise
deployment. Existing local web service • ArcGIS Pro can be used by GIS
transaction rates were increased by 10 Desktop clients to create and share
percent to accommodate projected system Portal content.
of engagement Portal for ArcGIS feature • Named users within the Portal
services. organization can create content and
publish new web maps.
The ArcGIS Enterprise deployment will • User needs include new Portal feature
significantly expand existing capabilities. services shared internally from Portal
content.
• Deploy Portal for ArcGIS for improved
security and establish an enterprise Best practice: ArcGIS Enterprise
system of engagement. deployment rapidly expands system
• Enable business users to create web capabilities with minimum deployment
maps and build configurable web apps risk.
from existing business resources.
• Leverage existing software licensing to
significantly expand operational
capabilities.
City of Rome 11.10

Workflow performance targets: ArcGIS Enterprise IOC
Figure 11.13 Workflow performance targets for the City of Rome ArcGIS Enterprise initial operational capability.
Figure 11.13 shows the workflow The workflow definition provides a recipe
performance targets defined for the ArcGIS that identifies the software technology
Enterprise IOC deployment. The CPT can performance factors used to generate each
be used to generate appropriate workflow user workflow performance target.
performance targets for the system design.
Best practice: User workflow
The ArcGIS Enterprise deployment includes performance targets are used by the CPT
new Portal workflows. to generate traffic and processing loads
on the CPT system design solution.
• DeskProMed. ArcGIS Pro desktop use
case (medium complexity) that will be
used by GIS viewers to expand display System design process: ArcGIS
and analysis capabilities for existing Enterprise IOC
user workflows. ArcGIS Pro will enable
GIS users to create new Portal content
and share feature services from existing
business resources.
• PortalFsvc. Portal named users can
create new web maps and apps and
share new feature services.
Performance target for these new
feature services is heavy complexity
with 20% dynamic layers (sharing
limited number of new business layers
over vector tile basemaps).
CPT Workflow Loads Analysis: ArcGIS Enterprise

IOC Figure 11.14 System design process provides a
logical step-by-step methodology for using the CPT to
CPT project workflows: ArcGIS Enterprise IOC complete your system design.
City of Rome 11.11

Figure 11.14 shows the process used to impacts will be shared following the
complete the system design. The system software configuration.
design process will be used to evaluate the • User requirements analysis.
ArcGIS Enterprise IOC deployment. • Network suitability analysis.
• Platform architecture selection.
System design process: • Software configuration.
• Enterprise design solution.
• Technical architecture strategy. ArcGIS
Enterprise IOC technical architecture
will be the same user locations and CPT Design requirements analysis:
network connectivity as the existing City ArcGIS Enterprise IOC
of Rome infrastructure. Network traffic
upgrades will impact technical CPT Design network suitability analysis:
architecture, and a summary of these ArcGIS Enterprise IOC
Platform architecture selection: ArcGIS Enterprise IOC
Figure 11.15 Platform architecture for ArcGIS Enterprise IOC deployment.
Figure 11.15 shows City of Rome ArcGIS Enterprise platform architecture components. Each
server machine will be supported by a high-availability virtual server platform architecture.
City of Rome 11.12

The platform architecture for the ArcGIS • DBMS tier: A total of six (6) DBMS
Enterprise IOC includes the following virtual servers (primary production and
machines. publication server machines, and a new
relational data store machine, each with
• WTS tier: ArcGIS Desktop ArcMap dedicated failover machines to satisfy
remote clients supported by XenApp high availability requirements).
sessions on Citrix virtual machines.
• Web tier: A total of six (6) web servers Best practice: A virtual server data
(internal and public web servers and a center environment provides optimum
new Portal server, each with two flexibility for deploying and supporting
machines) support the high-availability GIS operations.
Web tier architecture.
• GIS tier: A total of six (6) ArcGIS CPT Design platform configuration:
Servers (separate internal and public ArcGIS Enterprise IOC
GIS Server sites and a new internal
hosting server site, each site with two CPT Design software configuration:
machines to support high-availability ArcGIS Enterprise IOC
requirements).
Workflow performance summary: ArcGIS Enterprise IOC
Figure 11.16 Workflow performance summary demonstrates need for bandwidth upgrades to support ArcGIS
Enterprise IOC deployment.
Figure 11.16 shows the CPT Design Warning: ArcGIS Pro is not supported in
workflow performance summary for the a Citrix XenApp session terminal
ArcGIS IOC operations. Graphics show services environment.
before and after implementing the
recommended bandwidth upgrades.
City of Rome 11.13

ArcGIS Enterprise deployment with the • ArcGIS Pro operations over higher
existing network bandwidth will result in latency connections increase response
poor performance at the remote locations. time delays.
• Remote ArcGIS Desktop viewers will Best practice: The CPT Design network
need to use a local workstation install of suitability analysis can be used to
ArcGIS Pro. identify appropriate network upgrade
• ArcGIS Pro feature services generate requirements.
more traffic over the remote bandwidth
connections.
Technical architecture: ArcGIS Enterprise IOC network upgrades
Figure 11.17 City of Rome recommended bandwidth for ArcGIS Enterprise IOC deployment.
Figure 11.17 shows a summary of the City o Site 3 Freeberg—upgrade 45

of Rome ArcGIS Enterprise IOC Mbps to 155 Mbps WAN
recommended network upgrades. Network connection
bandwidth upgrades are required to support o Site 4 Willsberg—upgrade 45
ArcGIS Enterprise IOC remote user Mbps to 135 Mbps WAN
workflow performance requirements. connection
• Data center—45 Mbps Internet
Recommended network bandwidth connection (no change)
upgrades o Public web services will connect
through the data center Internet
• Data center—1000 Mbps LAN connection.
connection (no change)
• Data center—upgrade 90 Mbps to 310 Projected network cost (WAN/Internet
Mbps WAN connection Service) increases from $7,836/ month to
o Site 2 Operations facility— $12,289/month.
upgrade 12 Mbps to 24 Mbps
WAN connection Warning: The estimated network pricing
used in the CPT is for demonstration
purposes only.
City of Rome 11.14

Enterprise design solution: ArcGIS Enterprise IOC
Figure 11.18 City of Rome ArcGIS Enterprise IOC high-availability virtual server platform solution.
Figure 11.18 shows the ArcGIS Enterprise Server host platform load increased slightly
enterprise design solution. ArcGIS (31.6% to 34.3%) to accommodate
Enterprise IOC deployment replaces remote additional Portal for ArcGIS, ArcGIS Server
ArcGIS Desktop ArcMap viewers with an host server, and relational data store
ArcGIS Pro workstation deployment. processing loads. These additional loads
also include accommodation for the
CPT Design platform solution: ArcGIS Enterprise additional hosted feature services included
IOC in the user needs projection.
The following host platform tier support the The physical host server environment is
data center processing loads. supported by three (3) existing 2011 Xeon
X5687 8-core platforms.
• Citrix host platform: Not supported for
ArcGIS Pro clients. Best practice: ArcGIS Enterprise IOC
• Network bandwidth upgraded to support deployment has minimum impact on the
ArcGIS Pro existing host server processing loads.
workstation access over WAN from remote

locations.
• Server host platform: Three platform

nodes at 34.3% utilization.
City of Rome 11.15

Phase 1 Vector tile • Faster tile generation. World vector tiles
generated in 8 hours versus world raster
basemap deployment tiles generated in many weeks.
• Reduced tile size: World vector tiles ~13
The ArcGIS Enterprise deployment will GB; world raster tiles ~20 TB.
enable ArcGIS Desktop users to leverage
ArcGIS Pro capabilities to improve For City of Rome, vector tile basemaps for
operational effectiveness. City of Rome the operational area could be generated in
plans to allow ArcGIS Desktop viewers to less than 1 hour with tile size less than 1
leverage ArcGIS Pro and their legacy GB. Basemaps could be updated on a
ArcMap applications. ArcGIS Pro is not weekly basis to provide optimum operations
supported with the existing ArcGIS Desktop support.
terminal services environment. Existing
remote ArcGIS Desktop users must install Vector tiles are published and consumed by
ArcGIS Pro on their local workstations and ArcGIS client software.
access data sources over the existing WAN
environment. Vector tile basemaps can be • Tile creation: ArcGIS Pro 1.2+
used by ArcGIS Pro and local Web (geoprocessing services).
applications to reduce bandwidth costs and • Published tile layers: ArcGIS Online;
improve display performance for remote ArcGIS Server/Portal for ArcGIS 10.4+.
users over the WAN. • Client consumption: ArcGIS Runtime
Quartz; ArcGIS Pro 1.3+, ArcGIS
Vector tile basemaps JavaScript 3.15+, and 4.0 APIs.
Best practice: Vector tiles provide a

single layer of files that can be
dynamically displayed and styled.
For City of Rome, all existing Web apps and

ArcGIS Pro clients would be able to
leverage vector tile basemaps, significantly
reducing network traffic impacts, improve
Figure 11.19 Vector tile basemaps provide high- client display performance, and significantly
quality displays while minimizing traffic and improve capacity of existing servers.
processing loads.
Figure 11.19 shows a summary of the costs Workflow loads analysis: ArcGIS
and benefits of using vector tile basemaps. Enterprise IOC with vector tile
basemaps
Vector tiles are created and delivered with
much less processing. ArcGIS Enterprise IOC workflow loads
analysis will remain the same with vector tile
basemaps.
City of Rome 11.16

Workflow performance targets: ArcGIS Enterprise IOC with vector tile basemaps
Figure 11.20 Workflow performance targets for the City of Rome ArcGIS Enterprise IOC deployment with vector
basemaps.
Figure 11.20 shows the workflow The workflow definition provides a recipe
performance targets defined for the ArcGIS that identifies the software technology
Enterprise IOC deployment with vector performance factors used to generate each
basemaps. The CPT can be used to user workflow performance target.
generate appropriate workflow performance
targets for the system design. Best practice: Workflow performance
targets are used by the CPT to generate
ArcGIS Pro and the web services can traffic and processing loads on the CPT
leverage vector tile basemaps. Vector tile system design solution.
basemaps reduce display complexity by 60
percent. CPT Workflow Loads Analysis: Vector
tile basemap deployment
Warning: ArcMap is not able to use
vector tile basemaps. CPT project workflows: Vector tile
basemap deployment
New Portal workflows for ArcGIS Enterprise
IOC with vector tile basemaps.
• DeskProMed40%. DeskProMed
workflow with cached vector tile
basemap.
• WebInHvy40%. WebInHvy workflow
with cached vector tile basemap.
• WebPubMed40%. WebPubMed
workflow with cached vector tile
basemap.
City of Rome 11.17

System design process: ArcGIS System design process:
Enterprise IOC with vector tile
basemaps • Technical architecture strategy. ArcGIS
basemaps technical architecture will be
the same user locations and network
connectivity as the existing City of Rome
infrastructure. Network traffic
adjustments will impact technical
architecture and infrastructure costs,
and a summary of these impacts will be
shared following the software
configuration.
• User requirements analysis.
• Network suitability analysis.
• Software configuration.
Figure 11.21 System design process provides a • Enterprise design solution.
CPT Design requirements analysis:
Vector tile basemap deployment
complete the system design. The system CPT Design network suitability analysis:
design process will be used to evaluate the Vector tile basemap deployment
ArcGIS Enterprise IOC deployment.
City of Rome 11.18

Platform architecture selection: ArcGIS Enterprise IOC with vector tile basemaps
Figure 11.22 Platform architecture for ArcGIS Enterprise IOC deployment with vector tile basemaps.
Figure 11.22 shows City of Rome ArcGIS Negligible Web and GIS Server loads are
Enterprise IOC platform architecture required to support the client display tile
components with vector tile basemaps. cache downloads.
Each server machine will be supported by a
high-availability virtual server platform Best practice: A virtual server data
architecture. center environment provides optimum
flexibility for deploying and supporting
The ArcGIS Enterprise IOC virtual platform GIS operations.
configuration does not change. Vector tile
basemaps will be deployed on a data center CPT Design software configuration:
file share. Tiles will be downloaded to the Vector tile basemap deployment
client browser cache and displayed with the
local application as required to support the
display.
City of Rome 11.19

Workflow performance summary: ArcGIS Enterprise IOC with vector tile basemaps
Figure 11.23 Workflow performance summary demonstrates display performance improvements with vector tile
basemaps.
Figure 11.23 shows the CPT Design • Vector basemaps are downloaded once
workflow performance summary for the to the client browser cache.
ArcGIS IOC operations with vector tile • Dynamic processing loads are reduced
basemaps. Graphics show network cost due to the basemap cached layers.
savings and client display performance • Latency impacts are reduced due to less
improvements with the vector basemap dynamic layers (less chatter) in the
deployment. display.
Best practice: Vector tile basemaps Best practice: The CPT Design network
significantly reduce network traffic and suitability analysis can be used to
improve display performance. identify appropriate network upgrade
requirements.
ArcGIS Enterprise deployment with vector
basemaps reduces network traffic and
improves display performance.
City of Rome 11.20

Technical architecture: ArcGIS Enterprise IOC with vector tile basemaps
Figure 11.24 shows a summary of the o Site 3 Freeberg—reduce 155

benefits of deploying ArcGIS Pro and web Mbps to 90 Mbps WAN
services utilizing vector tile basemaps. connection
Network traffic requirements are reduced for o Site 4 Willsberg—reduce 155
the ArcGIS Enterprise IOC deployment Mbps to 90 Mbps WAN
when leveraging vector tile basemaps. connection
• Data center—reduce 45 Mbps to 24
Recommended network bandwidth savings Mbps Internet connection
o Public web services will connect
• Data center—1000 Mbps LAN through the data center Internet
connection (no change) connection.
• Data center—reduce 310 Mbps to 135
Mbps WAN connection Projected network cost (WAN/Internet
o Site 2 Operations facility— Service) decreases from $12,289/month to
reduce 24 Mbps to 12 Mbps $9,570/month.
WAN connection
Warning: CPT network pricing is used
for demonstration purposes only.
City of Rome 11.21

Enterprise design solution: ArcGIS Enterprise IOC with vector basemaps
Figure 11.25 City of Rome ArcGIS Enterprise IOC high-availability virtual server platform solution with vector
basemaps.
Figure 11.25 shows the ArcGIS Enterprise • Server host platform: Two (2) platform
IOC enterprise design solution with vector nodes at 30.3% utilization.
basemaps. ArcGIS Enterprise IOC
deployment replaces remote ArcGIS Server host platform load is reduced
Desktop ArcMap viewers with an ArcGIS significantly (3 platforms at 34.3% utilization
Pro workstation deployment. Vector to 2 platforms at 30.4% utilization)
basemaps reduce host server processing supporting the same ArcGIS Enterprise IOC
loads. workflows.
CPT Design platform solution: Vector tile The physical host server environment is
basemap deployment
supported by two (2) existing 2011 Xeon
X5687 8-core platforms.
The following host platform tier support the
data center processing loads.
Best practice: ArcGIS Enterprise IOC
deployment has minimum impact on the
• Citrix host platform: Not supported for existing host server processing loads.
ArcGIS Pro clients.
• Network traffic reduced from the initial
IOC design due to leveraging vector tile
basemaps.
City of Rome 11.22

Phase 1 Upgrade WTS host Figure 11.26 shows ArcGIS Pro display
performance when supported in a local
platform tier to support workstation or host server remote desktop
remote ArcGIS Pro clients environment. GRID GPU graphics must be
included with the host server configuration
to support ArcGIS Pro display performance.
ArcGIS Pro can be supported in a hosted
virtual desktop environment, which would
improve remote ArcGIS Pro display Host platform must include Tesla M60 GPU
performance and further reduce network graphics hardware and NVIDIA GRID
cost. Upgrading the WTS host platform tier concurrent user licensing
to support ArcGIS Pro clients would require
new hardware purchases. • NVIDIA Tesla M60 GPU
• NVIDIA GRID licensing for remote
desktop sessions (up to 25 concurrent
ArcGIS Pro remote desktop host
workstation clients)
platform requirements
ArcGIS Pro reference sites were provided in
the GIS Product Architecture chapter.
Warning: ArcGIS Desktop Citrix

XenDesktop deployment with raster
imagery included in the display requires
much higher client traffic. Pro 3D remote
client displays can experience
significantly more client traffic than what
Figure 11.26 ArcGIS Pro remote desktop is experienced with 2D displays.
deployment.
City of Rome 11.23

Workflow loads analysis: ArcGIS Pro Citrix remote clients
Figure 11.27 Citrix host platform upgrade to support remote ArcGIS Pro desktop users.
Figure 11.27 shows the remote viewers that • Remote DeskMed clients currently
would be hosted on the new ArcGIS Pro use ArcMap applications.
virtual desktop server. Display performance • ArcGIS Enterprise workflows will
for the remote ArcGIS Pro desktop users expand capabilities to use ArcGIS
can be improved by upgrading the Citrix Pro.
host platform tier with NVIDIA GRID • Existing ArcMap client licenses
graphics. include authorization to use ArcGIS
Pro.
Business requirements identify up to 52
remote ArcGIS Pro clients.
City of Rome 11.24

Workflow performance targets: ArcGIS Pro Citrix remote clients
Figure 11.28 Workflow performance targets for the City of Rome ArcGIS Enterprise IOC with vector basemaps and
ArcGIS Pro remote desktop.
Figure 11.28 shows Workflow performance Best practice: The user needs workflow
targets defined for the ArcGIS Enterprise loads analysis summary and the user
IOC deployment with vector basemaps and workflow performance targets are used
ArcGIS Pro remote desktop clients. The by the CPT to generate traffic and
CPT can be used to generate appropriate processing loads on the CPT system
workflow performance targets for the design solution.
system design.
CPT Workflow Loads Analysis: ArcGIS Pro Citrix
ArcGIS Pro remote desktop workflow (40% remote clients
dynamic) will leverage the vector tile
CPT project workflows: ArcGIS Pro Citrix remote
basemaps. clients deployment
The ArcGIS Enterprise IOC with vector

System design process: ArcGIS Pro
basemaps includes new ArcGIS Pro
workflows. Citrix remote clients
• CitrixProMed40%. Remote ArcGIS Pro

clients using host Citrix remote desktop
server tier and vector basemaps.
• CitrixBatch. ArcGIS Pro workflows
include a variety of geoprocessing
capabilities that users can leverage
within the remote desktop workflows.
CitrixBatch workflow provides batch
loads on the Citrix server for each
geoprocessing instance.
The workflow definition provides a recipe

that identifies the software technology Figure 11.29 System design process provides a
performance factors used to generate each logical step-by-step methodology for using the CPT to
user workflow performance target.
City of Rome 11.25

Figure 11.29 shows the process used to and a summary of these impacts will be
complete the system design. The system shared following the software
design process will be used to evaluate the configuration.
ArcGIS Enterprise IOC deployment. • User requirements analysis.
System design process: • Platform architecture selection.
• Technical architecture strategy. ArcGIS • Enterprise design solution.
basemaps and remote ArcGIS Pro CPT Design requirements analysis:
virtual desktops technical architecture ArcGIS Pro Citrix remote clients
will be the same user locations and deployment
network connectivity as the existing City
of Rome infrastructure. Network traffic CPT Design network suitability analysis:
adjustments will impact technical ArcGIS Pro Citrix remote clients
architecture and infrastructure costs, deployment
Platform architecture selection: ArcGIS Enterprise IOC
Figure 11.30 Platform architecture with vector tile basemaps and option for ArcGIS Pro remote desktop clients.
Figure 11.30 shows City of Rome ArcGIS clients. Each server machine will be
Enterprise IOC platform architecture supported by a high-availability virtual
components with vector tile basemaps and server platform architecture.
support for ArcGIS Pro remote desktop
City of Rome 11.26

The technical platform architecture for the pricing for City of Rome procurement. City
ArcGIS Enterprise IOC with vector tile of Rome management requests that you
basemaps and ArcGIS Pro remote desktops complete a cost analysis for the various
include the following additional virtual platform architecture patterns being
machines. proposed for this study. For this analysis,
City of Rome provided their hardware price
• WTS tier: ArcGIS Desktop ArcMap and list for platforms under consideration for this
ArcGIS Pro remote desktop clients. design.
Best practice: A virtual server data Esri recommends a 28-core server platform
center environment provides optimum with NVIDIA GRID graphics to support 25
flexibility for deploying and supporting concurrent ArcGIS Pro clients. High-
GIS operations. capacity server is required to support the
batch geoprocessing jobs performed by
clients during typical ArcGIS Pro workflows.
Geoprocessing needs will determine server
capacity requirements.
City of Rome hardware price list
Warning: Customer price lists can vary
based on vendor arrangements and
contract agreements. It is important to
validate pricing and update platform
pricing on the CPT Hardware Pricing tab
if you want to include pricing in your
analysis.
CPT Design platform configuration:

ArcGIS Pro Citrix remote clients
deployment
Figure 11.31 City of Rome hardware price list is used CPT Design software configuration:
to complete this business case analysis. ArcGIS Pro Citrix remote clients
deployment
Figure 11.31 shows the 2017 host hardware
platform options and associated vendor
City of Rome 11.27

Workflow performance summary: ArcGIS Pro Citrix remote clients
Figure 11.32 Display performance gain with ArcGIS Pro remote desktop clients.
Figure 11.32 shows display performance • ArcGIS Pro clients have display and
summary for the ArcGIS Enterprise IOC control of the application from their
with ArcGIS Pro remote desktop workflows. remote site location.
• Network traffic is reduced because
Best practice: ArcGIS Pro remote only the final display is sent to the
desktop performance is a significant client.
improvement over remote ArcGIS Pro • Latency impacts are reduced due to
workstation clients accessing the less chatter with the remote desktop
database over the WAN network. display.
ArcGIS Enterprise deployment with ArcGIS Best practice: The CPT Design network
Pro remote desktop clients improves display suitability analysis can be used to
performance. identify appropriate network upgrade
requirements.
• ArcGIS Pro application is supported
in the data center with local data
sources.
City of Rome 11.28

Technical architecture: ArcGIS Pro Citrix remote desktop traffic improvements
Figure 11.33 shows a summary of the o Site 4 Willsberg—reduce 90

benefits of ArcGIS Pro remote desktop Mbps to 24 Mbps WAN
deployment. Network traffic requirements connection
are reduced for the ArcGIS Enterprise IOC • Data center—24 Mbps Internet
deployment when leveraging ArcGIS Pro connection (no change)
remote desktops. o Public web services will connect
through the data center Internet
Recommended network bandwidth savings connection
• Data center—1000 Mbps LAN Projected network cost (WAN/Internet

connection (no change) Service) decreases from $9,570 /month to
• Data center—reduce 135 Mbps to 45 $5,986/month.
Mbps WAN connection
o Site 2 Operations facility—12 Warning: CPT network pricing is used
Mbps WAN connection (no for demonstration purposes only.
change)
o Site 3 Freeberg—reduce 90
Mbps to 24 Mbps WAN
connection
City of Rome 11.29

Enterprise design solution: ArcGIS Pro Citrix remote clients
Figure 11.34 City of Rome ArcGIS Enterprise IOC solution with ArcGIS Pro remote desktops.
Figure 11.34 shows a mix of 52 concurrent • Additional Citrix host platform can be
ArcMap and ArcGIS Pro remote desktop added at a later date as user migration
clients, with no more than 50 concurrent to ArcGIS Pro dictates.
ArcGIS Pro clients.
CPT Design platform solution: ArcGIS Pro Citrix
The Citrix host platform remote clients deployment
• New E5-2690v4 28-core server with two The Server host platform
NVIDIA GRID Tesla M60 video cards.
• Citrix tier solution on the left (5.4% • 2011 X5687 8-core server.
utilization) does not include any ArcGIS • Two (2) platform nodes at 30.4%
Pro geoprocessing jobs. utilization.
• Citrix tier solution on the right (74.5%
utilization) includes 26 concurrent batch Best practice: ArcGIS Pro background
jobs. geoprocessing needs dictate host
platform capacity.
City of Rome 11.30

Phase 1 ArcGIS Enterprise IOC baseline design solution
Figure 11.35 City of Rome ArcGIS Enterprise IOC hardware platform upgrade baseline.
Figure 11.35 shows an ArcGIS Enterprise Pro geoprocessing jobs. Column AF

IOC hardware upgrade baseline solution identifies the capability to support up to
(includes 2017 Server host platform 26 concurrent batch jobs in conjunction
upgrade). with the client loads.
• Citrix tier solution on the right (74.5%
CPT Phase 1 ArcGIS Enterprise IOC utilization) includes 26 concurrent batch
baseline design solution jobs.
The Citrix host platform (upgraded in the The Server host platform (new 2017
previous slide) hardware)
• New E5-2690v4 28-core server with two • E5-2637v4 8-core server.

NVIDIA GRID Tesla M60 video cards. • Two (2) platform nodes at 22.1%
• Citrix tier solution on the left (5.4% utilization.
utilization) does not include any ArcGIS
City of Rome 11.31

Phase 2 Water Utility Solution
For Phase 2, City of Rome plans to deploy the ArcGIS Water Utility Solution with templates
available on the Esri Web site.
Figure 11.36 ArcGIS Solutions for Water Utilities.
Figure 11.36 shows the ArcGIS Water • ArcGIS Desktop - 11

Solutions template. ArcGIS Solutions • Collector for ArcGIS - 13
include ready-to-use maps and apps to • ArcGIS GeoEvent Server - 4
jumpstart your work. Phase 2 leverages • Operations Dashboard for ArcGIS - 4
ArcGIS Solutions to implement an upgrade • Web AppBuilder for ArcGIS - 8
of the City of Rome water utility operations. • Web Applications - 20
The ArcGIS Water Solutions include over 46 Solutions also include a model organization
maps and apps that can be downloaded template for ArcGIS Online and Portal for
and used by customers to jumpstart their ArcGIS.
GIS operations. Solutions include water-
specific information products supported by These ArcGIS solutions are samples used
the following applications: for typical Water Delivery, Sewer Collection,
and Stormwater Conveyance operations.
City of Rome 11.32

Technical architecture: Phase 2 Operations and Water Utilities existing operations
Figure 11.37 City of Rome existing technical architecture and network connectivity following Phase 1 deployment.
Figure 11.37 shows a summary of the o Site 9 Rogerton—1.5 Mbps WAN

Phase 2 user locations and existing network connection
connectivity. • Data center—24Mbps Internet
connection
City Hall data center remote network o Public services connect through
connections (Phase 1 baseline) are as Internet connection
follows:
Reference network cost ($8,935/month)
connection Workflow loads analysis: Water
• Data center—45 Mbps WAN connection Utilities solution
o Site 2 Operations facility—12
Mbps WAN connection The workflow loads analysis identifies the
o Site 3 Freeberg—24 Mbps WAN workflow patterns supporting user business
connection requirements, establishes a display
o Site 4 Willsberg—24 Mbps WAN complexity (performance target) for each
connection workflow, and identifies user locations (local
o Site 5 Perth—1.5 Mbps WAN or remote) and peak concurrent system
connection loads (users or transaction rates) for each
o Site 6 Wawash—1.5 Mbps WAN workflow pattern.
connection
o Site 7 Jackson—1.5 Mbps WAN
connection
o Site 8 Petersville—1.5 Mbps
WAN connection
City of Rome 11.33

Workflow display complexity: Water Display complexity should be reviewed and
Utilities solution updated for the City of Rome Geodatabase.
• Review Water Solution map

performance with City of Rome
Geodatabase.
• Adjust complexity estimates for
remaining Water Solutions based on
measured performance.
User workflow complexity is an estimate of

the processing loads required for an
Figure 11.38 User workflow patterns are normally a average information product display
product of a user-needs assessment and provide a transaction. The workflow complexity
reference for establishing workflow loads for the estimate will be used with the CPT to
system design.
generate appropriate system design
performance specifications.
Figure 11.38 shows workflow patterns
identified for the Water Solutions. Display
Best practice: A conservative workflow
complexity is identified for each of the Water
complexity will ensure an adequate
maps and apps, and workflows are grouped
system design solution.
by technology pattern and display
complexity.
Water user needs template
Figure 11.39 Water user needs template modified for City of Rome planning.
Figure 11.39 shows a Water Solutions user needs template that can be used as a framework for
gathering business requirements.
City of Rome 11.34

Template groups ArcGIS Water solutions The Water user needs template is modified
based on relative display complexity. to show the City of Rome user needs.
• Desktop: medium, heavy, 2xmedium, • Location along with department and

and batch processing complexity network connectivity (Columns A-D).
• Collector: light and medium processing • Total Water maps and apps users
complexity (column E).
• Dashboard: light and medium • Peak user loads for each application
complexity category (column F-CT).
• Web Apps: light, medium, heavy,
2xmedium, and GeoEvent feeds Best practice: Identify peak load
• GeoEvent Server: events per second estimates by software technology and
complexity.
Best practice: A Water user needs
template is used for workflow loads
analysis.
Phase 2 Operations and Water Utilities user needs
Figure 11.40 Phase 2 Operations and Water Utilities user needs summary.
Figure 11.40 shows results of the City of The Phase 2 deployment will initially
Rome Phase 2 user needs assessment. leverage the Water Solutions maps and
Results of the Water user needs analysis apps.
provides information needed to complete a
composite user needs summary for the • Delivers rapid deployment with COTS
Phase 2 Operations and Water Utilities maps and apps.
deployment. • Minimizes deployment risk.
City of Rome 11.35

Water maps and apps can be expanding to • Water Solutions performance categories
support custom business requirements. support typical Water deployments.
• System design is based on generic Best practice: User needs analysis and
performance targets. system design based on the Water
• Custom Water applications can be Solutions templates reduces
deployed in same categories. implementation risk and accelerates
return on investment.
Workflow performance targets: Water Utility deployment
Figure 11.41 Workflow performance targets for the Phase 2 City of Rome GIS operations.
Figure 11.41 shows the workflow New ArcGIS Desktop workflows

performance targets selected for the City of
Rome Phase 2 deployment. The CPT can • DeskMap2Med. ArcMap workstation
be used to generate appropriate workflow 2xMedium workflows.
performance targets for the system design. • DeskPro2Med40%. Pro workstation
2xMedium workflows.
Phase 2 builds on the Phase 1 deployment. • DeskGP. Desktop workstation
geoprocessing services.
• CitrixMap2Med. ArcMap remote desktop
2xMedium workflows.
City of Rome 11.36

• CitrixPro2Med40%. Pro remote desktop System design process: Phase 2
2xMedium workflows. Water Utilities deployment
• CitrixGP. ArcMap and Pro remote
desktop batch services.
New Collector, GeoEvent, and Operations

Dashboard workflows
• MobileMed10%. Collector for ArcGIS

mobile client workflow.
• GeoEventMed. GeoEvent streaming
services.
• DashMed40%. Operations Dashboard
clients.
New Web Application workflows

Figure 11.42 System design process provides a
• Webin2Med40%. New Webinternal complete your system design.
2xMedium workflows.
CPT Workflow Loads Analysis: Water complete the system design. The system
Utility deployment design process will be used to evaluate the
Phase 2 Water Utilities deployment.
CPT project workflows: Water Utility
deployment System design process:
• Technical architecture strategy. Figure

37 shows City of Rome Phase 2
technical architecture strategy. Network
traffic adjustments will impact technical
architecture and infrastructure costs,
and a summary of these impacts will be
shared following the software
configuration.
• User requirements analysis.
• Enterprise design solution.
City of Rome 11.37

Platform architecture selection: Phase 2 Water Utilities deployment
Figure 11.43 GIS data center platform architecture for Phase 2 City of Rome GIS Operations and Water Utilities
deployment.
Figure 11.43 shows The Phase 2 and external relational data store,
Operations and Water Utilities platform and public geospatial data store.
architecture.
Each platform environment must be
The Phase 2 Operations and Water Utilities configured to support high-availability
platform architecture includes the following. requirements.
• WTS tier: ArcGIS Desktop ArcMap Best practice: A virtual server data
and Pro remote desktop clients. center environment provides optimum
• Web tier: Internal and public web flexibility for deploying and supporting
and Portal servers. GIS operations.
• GIS tier: Internal and public GIS
publication and hosting server sites, CPT Design requirements analysis:
plus GeoEvent server machines. Water Utility deployment
• DBMS tier: Production and
Publication geodatabase, internal CPT Design network suitability analysis:
Water Utility deployment
City of Rome 11.38

Workflow performance summary: Phase 2 Water Utilities deployment
Figure 11.44 Workflow performance summary for the Phase 2 City of Rome business workflows.
Figure 11.44 shows the CPT Design Phase Important to note that the CPT model shows
2 workflow performance summary. average performance levels.
Workflow performance summary provides • Some display response times can be

the following representation. lighter.
• Other display response times will be
• Average processing and queue times for slower.
each user workflow. • Queue times will vary depending on
• Average difference in performance random load profiles.
between each site location.
• Average performance for shared public Best practice: Workflow performance
services. summary can be used to identify realistic
performance expectations.
City of Rome 11.39

Technical architecture: Phase 2 Water Utilities deployment (network bandwidth
upgrades)
Figure 11.45 City of Rome recommended bandwidth for Phase 2 Water Utilities deployment.
Figure 11.45 shows a summary of the o Site 8 Petersville WAN—

Phase 2 user locations and network upgrade 1.5 Mbps to 45 Mbps
connectivity. Network traffic will require o Site 9 Rogerton WAN—upgrade
bandwidth upgrades to support the 1.5 Mbps to 45 Mbps
identified Phase 2 user requirements. • Data center Internet—upgrade 24 Mbps
to 2500 Mbps
Phase 2 network upgrades. o Public services connect through
Internet connection
connection (no change) Current network cost (WAN/Internet
• Data center WAN—upgrade 45 Mbps to Service) increases from $8,935/ month to
310 Mbps $20,057/month.
o Site 2 Operations facility WAN—
upgrade 12 Mbps to 135 Mbps Warning: CPT network pricing is used
o Site 3 Freeberg WAN—upgrade for demonstration purposes only.
24 Mbps to 45 Mbps
o Site 4 Willsberg WAN—upgrade CPT Design generic 4-tier platform
24 Mbps to 45 Mbps configuration: Water Utility deployment
o Site 5 Perth WAN—upgrade 1.5
Mbps to 12 Mbps CPT Design generic 4-tier software
o Site 6 Wawash WAN—upgrade configuration: Water Utility deployment
1.5 Mbps to 24 Mbps
o Site 7 Jackson WAN—upgrade CPT Design generic 4-tier platform
1.5 Mbps to 18 Mbps solution: Water Utility deployment
City of Rome 11.40

Enterprise design solution: Phase 2 Water Utilities deployment
Figure 11.46 City of Rome ArcGIS Phase 2 Operations and Water Utilities deployment.
Figure 11.46 shows the recommended • Capacity to support up to 100

platform solution for the City of Rome Phase concurrent ArcGIS Pro background jobs.
2 Operations and Utilities deployment.
Server host platform: Xeon E5-2643v4 12-
The following host platform tier support the core servers
data center processing loads.
• Three platform nodes at 48.2%
The Citrix host platform: Xeon E5-2690v4 utilization.
28-core servers
Best practice: Proper system design
• 8 platform nodes at 10.3% utilization. promotes implementation success.
• Supports up to 209 concurrent ArcMap
and ArcGIS Pro remote desktop
sessions.
• Maximum of 200 ArcGIS Pro remote
desktop sessions (graphics card
limitation).
City of Rome 11.41

City of Rome system design cost analysis summary
Figure 11.47 City of Rome case study cost analysis
Figure 11.47 provides a final business cost • Phase 1 hardware upgrade baseline
summary supporting the available City of o Additional Server tier upgrade
Rome GIS deployment options. costs ($36,334 increase)
Phase 1 deployment options Phase 2 Water Utilities deployment
• ArcGIS Enterprise IOC deployment • Estimated hardware server cost=

o Operational cost increase due to $395,968
network traffic ($53,436 annual • Estimated virtualization pricing =
increase) $109,890/year
• Implement vector tile basemaps • Estimated GPU pricing = $160,000
o Operational cost savings due to • Estimated network pricing =
reduced network traffic ($19,152 $240,684/year
savings)
• Implement ArcGIS Pro remote desktop Warning: Pricing models in CPT are used
o Operational cost savings due to for demonstration purposes only and
reduced network traffic ($42,180 should be updated with actual costs for
savings) each customer system design analysis.
o Additional host server upgrade
costs (124,928 increase)
City of Rome 11.42

Choosing a system configuration
The best solution for a given organization depends on the distribution of the user community
and the type of operational data in use. User requirements determine the number of machines
necessary (to support the operational environment), the amount of memory required (to support
the applications), and the amount of disk space needed (to support the system solution). The
system design models provide target performance metrics to aid in capacity planning. The
capacity planning tool incorporates standard templates representing the sizing models and
provides a manageable interface to help in enterprise-level capacity planning efforts. The CPT
can be a big help in applying the results of the user needs assessment.
User needs change as organizations change, so this assessment not only identifies platform
and infrastructure specifications and sets performance targets for the initial implementation, it is
also part of the process going forward. System upgrades, new technology solutions, tuning and
optimizing performance--every implementation or change is like a new launch, insofar as you
need to plan for it. Planning provides an opportunity to establish performance milestones that
can be used to manage a successful GIS implementation. Performance targets used in capacity
planning can provide target milestones to validate performance and scalability throughput
deployment of the system.
City of Rome 11.43

System Implementation
Fall 2017 System Implementation 41st Edition
Successful system implementation requires good leadership and careful planning. A good
understanding of every component of the system is critical in putting together an implementation
strategy. Enterprise IT environments involve integration of a variety of vendor technologies.
Interoperability standards within commercial software environments are voluntary, and even the
simplest system upgrade must be validated at each step of the integration process.
Enterprise GIS environments include a system deployment are directly related to

broad spectrum of technology integration. the variety of vendor components required
Most environments today include a variety to support the final production solution.
of hardware vendor technologies including
database servers, storage area networks, Centralized computing solutions with a
Windows Terminal Servers, Web servers, single database environment are the easiest
map servers, and desktop clients, —all environments to implement and support.
connected by a broad range of local area Distributed computer systems with multiple
networks, wide area networks, and Internet distributed database environments can be
communications. All these technologies very complex and difficult to deploy and
must function together properly to support a support. Many organizations are
balanced computing environment. consolidating their data resources and
application processing environments to
A host of software vendor technologies reduce implementation risk and improve
including database management systems, administrative support for enterprise
ArcGIS Desktop and ArcGIS Server business environments.
software, Web services, and hardware
operating systems—all integrated with
existing legacy applications. Data (including
GIS Staffing
business layers, basemap layers, and
Imagery) and user applications are added to People are the most valuable asset to any
the integrated infrastructure environment to organization. GIS managers have the
support the final implementation. The result special opportunity and challenge to bring
is a very large mixed bag of technology that people across their organization together to
must work together properly and efficiently share responsibility for building and
to support user workflow requirements. maintaining enterprise GIS operations. In
most cases, building a GIS is not a single
department responsibility – and in many
The integration and implementation of
cases the quality of the GIS depends on
distributed computer technology have
people outside the organization. GIS brings
become easier over the years as interface
people together because they depend on
standards have matured. At the same time,
each other in establishing the framework for
enterprise environments have become
building geographic information products
larger and more complex. The complexity
that support their business needs.
and risk associated with an enterprise
System Implementation 12.1

Identify Key Staff Functions
Figure 12.1 Key operational staffs participate in data management, planning and analysis, field mobility, and
operational awareness disciplines. Key supporting positions include skills in application development and enterprise
GIS management.
A broad range of technical skills are GIS Organizational Structure

required to build and maintain effective
enterprise GIS operations. Key operational
staffs participate in data management,
planning and analysis, field mobility, and
operational awareness disciplines. Key
supporting positions include skills in
application development and enterprise
system administration. Figure 12.1 provides
an overview of the variety of GIS workflows
Figure 12.2 GIS management must be organized to
and key functional responsibilities required facilitate implementation across multiple departments
to support enterprise GIS operations. and often extend to include a variety of business units
within a broader community of GIS users.
The complexity of these responsibilities will
vary with the size and extent of each Figure 12.2 shows an overview of a
individual GIS implementation, although traditional GIS matrix organization.
every organization will need some level of Enterprise GIS operations must be
support and expertise in each of these supported by an executive committee with
areas. influence and power to make financial and
policy decisions for the GIS user
Building Qualified Staff community. The GIS manager should
establish a coordinating committee
Training is one way to build qualified staff responsible for providing technical direction
and improve GIS user productivity. and leadership. Committee leaders should
Business organizations should develop and chair working groups assigned and aligned
maintain a comprehensive training plan to with each technical discipline to address
make sure their teams have the resources organizational issues and report on system
and skills they need to be effective in their status. The user community should be
job. represented throughout the review process.

A formal organizational structure provides a managing GIS in small to large
framework for establishing and maintaining organizations, and the same type of
the long-term support required for organizational structure can be effective in
successful enterprise GIS operations. This managing community GIS operations.
basic organization structure can be useful in
Integrated system design process
Figure 12.3 Integrated business needs assessment promotes proper and timely business system design decisions.
Traditionally, the user needs assessment and the system architecture design were two separate
efforts. There are some key advantages in completing these efforts together. Figure 12.3 shows
an overview of the system architecture design development methodology. Enterprise architects
now have tools they can use to complete an integrated business needs assessment.
There are four architecture domains that are commonly accepted as subsets of an overall
integrated business needs assessment. These include generally accepted guidelines and best
practices provided by The Open Group global consortium.
• The Business Architecture defines the business strategy, governance, organization, and
key business processes.
• The Information Systems Architecture includes a review of the Data and Application
architecture.
o The Data Architecture describes the structure of an organization’s logical and
physical data assets and data management resources.
o The Application Architecture provides a blueprint for the individual applications to
be deployed, their interactions, and their relationships to the core business
processes of the organization.

• The Technology Architecture describes the logical software and hardware capabilities that
are required to support the deployment of business, data, and application services. This
includes IT infrastructure, middleware, networks, communications, processing, standards,
etc.
Best practice: System architecture design should be included as an integral part of every
business planning process.
The GIS integrated business needs assessment (user needs/system architecture design)
provided in the SDSwiki documentation shares a tailored architecture development
methodology to promote successful Enterprise GIS operations. The primary steps for
completing an integrated design solution include the following:
• Business architecture
o Enterprise vision identifies how GIS technology can best support your business
needs.
• Information systems architecture
o Existing business architecture reviews your current data center operations to
identify existing experience for deployment and maintenance of available target
architecture options.
o Workflow loads analysis identifies business workflows and peak processing loads
that must be supported by the target architecture solution.
o Technical architecture strategy identifies data center and business user locations
and network connectivity for the proposed target architecture solution.
o User requirements analysis combines user location and workflow loads analysis to
identify distributed network and platform processing loads and user productivity.
• Technical architecture
o Network suitability analysis identifies peak network bandwidth requirements during
peak system loads.
o Platform architecture selection identifies data center platform configuration for the
target architecture solution.
o Software configuration identifies Enterprise level workflow software loads applied
to the data center target architecture selection.
o Enterprise design solution provides a summary of the final target architecture
solution that includes required platform nodes, projected platform utilization, peak
user workflow loads, and projected user productivity.
The SDSwiki Capacity Planning Tool provides a framework for completing the system
architecture design. Once the user requirements and architecture solution are configured, the
CPT completes the system architecture design loads analysis to identify network bandwidth and
the platform target state design solution.
Once you have completed your System Architecture Design and identified your target
architecture solution, suggested guidance for deploying Enterprise design solutions are
provided for each of the following implementation phases.
• Opportunities and Solutions identify the architecture deployment roadmap that will deliver
continuous business value.

• Migration planning identifies the optimum deployment strategy for maintaining existing
operations while upgrading operations to the target state.
• Implementation governance establishes oversight and monitoring required to manage
successful deployment migration to the target state.
• Architecture change control establishes procedures to maintain successful operations
through the migration to the final operational target state.
• Requirements management ensures procedures are in place and followed to successfully
execute the migration strategy.
The primary TOGAF Architecture Development Method (ADM) is described in Part-II of the
TOGAF®9.1 online documentation.
Best practice: The optimum deployment approach will depend on your specific
Enterprise design and business operations complexity.
Pre-design efforts familiar with current GIS technology

patterns and customer business practices
can help facilitate this planning effort; the
real work must be done by the organization
themselves.
Enterprise vision
Figure 12.4 Business needs establish the foundation

for any enterprise GIS design. The enterprise vision,
existing business architecture, and user requirements
must be understood to select the best GIS solution.
Figure 12.4 shows how you can prepare for

your system architecture design. Business
needs must be understood before you are
ready to complete the system architecture
design. Business requirements analysis Figure 12.5 GIS technology has evolved to support a
includes a review of the enterprise vision, broad integrated range of business needs across the
the existing business architecture, and the organization. Each GIS technology pattern is
user workflow requirements. Each of these optimized to address specific organizational needs.
areas must be explored in some detail
before you begin the design. Figure 12.5 shows an overview of the
ArcGIS technology patterns. GIS enterprise
The GIS needs assessment begins with the vision looks at how GIS technology can best
organization identifying where GIS support your business needs. ArcGIS
technology can improve the quality and includes a range of technology options
productivity of the business process flow. developed as a complete set of integrated
This assessment identifies GIS application workflows and systems to satisfy a broad
and data requirements and an range of business requirements.
implementation strategy for supporting GIS
user needs. The user organization must be GIS software deployment patterns are
actively involved throughout the user needs optimized to support your business needs:
assessment. A GIS solutions architect

• Asset management a design solution that can be supported best
• Planning and analysis by the organization.
• Field mobility
• Operational awareness Meet with the GIS and IT managers to
• Constituent engagement review any policies or preferences they may
have for the design. Major system upgrades
Most successful enterprise GIS operations often provide a chance for reviewing new
evolve to embrace the full range of available platform technology directions, and
GIS technology patterns to address focused management may want to include specific
business needs throughout their alternative technology patterns they are
organization. considering in the design effort.
Best practice: Establishing a clear Use the right language

enterprise GIS vision early in planning
can help identify an optimum roadmap It is important to recognize and use the
for building effective GIS operations. proper terminology when discussing design
issues.
Existing Business Architecture Best practice: Take time to listen.
Successful design consulting requires some

communication skills.
• Often the language used by the

technical staff is very different than what
is used by the user community.
• Performance and user productivity is
often viewed differently by the IT staff
and the user community.
• Technology is changing rapidly, along
Figure 12.6 Business architecture defines the current
state of how you are meeting your business
with the words that are used to describe
requirements. system loads, architecture patterns,
system performance and capacity
This includes a review of your platform and needs.
network architecture, governance and • Words that are used to describe GIS
political landscape, types of user technology patterns change and many
communities, operational constraints and design concepts are not well
priorities, and existing funding constraints. understood.
This is information that can be leveraged to
identify a GIS design solution that builds on Best practice: The words you use, how
your current business operations. you listen, and how you speak
establishes your credibility with both the
Governance and political landscape user community and the technical IT
staff. Credibility is very important in
Organizations often develop policies and leading a mixed group of business
standards that support their software and users, technical architects, and system
hardware investment decisions. A review of administrators toward a proper design
management preferences and associated decision.
vendor relationships will provide insight into

Platform and network environments Security requirements
The current IT environment can provide Identify the level of security governing the
insight into administrative staff experience current business operations
and policies in working with available
technology. Whether on your own or in • Basic - no sensitive data.
concert with a design consultant, you should • Standard - moderate consequences for
review the vendor platforms and network data loss or integrity
environments currently maintained by your • Advanced - sensitive data
organization. Hardware experience,
maintenance relationships, and staff training What security standards are currently in
represent a considerable amount of place?
investment for any organization.
• Published Web services standards
Best practice: Proposed GIS design • Data production and distribution access
solutions should take advantage of standards
corporate experience gained from • Access protection for Web application
working with the established platform servers and data sources.
and network environment.
Performance requirements
Operational constraints and priorities
Identify any performance concerns being
Understanding the type of operations addressed by the new design
supported by the GIS solution will identify
requirements for fault tolerance, security, • User productivity
application performance, and the type of • Remote access
client/server architecture that would be • Public web services
appropriate to support these operations. • Geoprocessing timelines
• Batch process timelines
System availability requirements
Best practice: High availability,
Most enterprise operations include several redundancy, security, and special
additional platform requirements in addition performance considerations drive
to their production environment. requirements for increased hardware
and software costs. Recommendations
• Development and test platforms should be backed up with facts to
• Staging platforms support proper cost and benefit analysis.
• Redundant maintenance and publishing
database environments Funding constraints
• Possible remote backup data center
• Possible cloud collaboration and Recommended solutions must fit within
publishing services reasonable organizational funding
constraints or they will not be accepted.
Warning: All business requirements and
priorities are not the same, and it is Warning: The final design must be
important to listen and understand what affordable.
is important in making the final design
recommendations. An organization will not implement a
solution that is beyond its financial
resources.

• With system design, cost is a function of • Managing the evolution of enterprise
performance and reliability. GIS operations.
• If cost is an issue, the system design
must facilitate a compromise between Technology is changing more rapidly every
user application performance, system year.
reliability, cost and schedule.
• The design consultant must identify a • During the 1990s, GIS planning was a
hardware solution that provides detailed, rigorous process required to
optimum performance and reliability identify and justify major changes in
within identified budget constraints. business processes necessary to
achieve the benefits provided by GIS
Current technology enables distribution of technology. GIS implementation would
GIS solutions to clients throughout an take several years to reach the final
enterprise environment, but there are planned state—and technology would
limitations that apply to any distributed be relatively stable throughout that
computer system design. period.
• Today, technology is changing much
• It is important to clearly understand real faster, and it is difficult to plan for more
GIS user needs and discuss alternative than one year at a time. Technology
options for meeting those needs with keeps improving, and adjustments must
system support staff to identify the most be made each year to keep pace with
cost-effective solution. the change. Planning methodology is
• It may be necessary to review several becoming more agile, adapting to the
alternative software technology patterns rapid change in technology.
along with a variety of system
deployment options to identify and Best practice: Enterprise GIS planning is
establish the best implementation an ongoing process, and should be
strategy. updated on an annual basis to keep pace
with technology.
Maintain a current plan
Figure 12.7 GIS planning should be updated annually

and integrated into your overall enterprise business
planning process.
Figure 12.7 emphasizes the importance of Figure 12.8 CPT can be very useful tool for
maintaining a current GIS plan. representing your enterprise GIS operations.
Planning is critical for: Most GIS deployments evolve over many

years of incremental technology
• Providing a framework for enterprise improvements, and:
GIS implementation.
• Ensuring upper management support for • Implementation plans normally address
required GIS investments. a two- or three-year schedule to ensure

that the budget is in place for the GIS enterprise operations are both compute
anticipated deployment needs. processing and network traffic intensive,
• Project planning should be adjusted which means:
annually to take advantage of
technology improvements and adjust for • GIS operations can place heavy
technology change. demands on server processing and
network traffic loads.
Figure 12.8 shows the CPT Design, a • Network capacity can be one of the
snapshot of a GIS plan identifying user determining factors driving proper
needs, site locations, peak system loads, software technology selection.
and platform solution for a specific target • In some cases, hardware constraints
architecture (point in time). Changes in user can drive software technology selection.
workflow requirements will adjust loads on
the selected platform solution. Changes in Best practice: Infrastructure
the platform architecture will identify requirements should always be
expected performance improvements and understood and considered before
system capacity. The CPT is a simple making a final software technology
analysis tool that is easy to change and selection.
understand. The CPT provides an adaptive
design model representing your GIS
enterprise environment.
System Design Process
Best practice: The Capacity Planning
Tool provides simple to use framework
for GIS system design analysis and
planning.
System architecture design
Figure 12.10 System Architecture Design process

provides a logical step by step methodology for using
the CPT to complete your System Architecture
Design.
Figure 12.9 System architecture design provides the Once you have identified your project
foundation for building a successful GIS. workflows, you are ready to complete your
system design. The CPT is developed for
Figure 12.9 shows the system architecture use based on a standard system
design process. System architecture design architecture design process as shown in
is an integral part of the GIS business Figure 12.10. Each cycle of the system
needs assessment. architecture design process includes the
following steps:

• Technical architecture strategy—High-
level overview showing user site System Architecture
locations, network bandwidth
connections, and central data center Deployment Strategy
locations. User location information is
collected during the user needs
analysis.
• User requirements analysis—CPT
Requirements analysis section is
configured to represent the site
locations, user workflows, peak loads,
and network bandwidth for the
enterprise design solution. Figure 12.11 A phased system deployment strategy
includes prototype development testing, initial
• Network suitability analysis—CPT production roll-out and final production roll-out on an
Design completes the network suitability established implementation schedule that enforces
analysis and identifies any system configuration control.
communication bottlenecks. Network
bandwidth upgrades are identified to Planning is the first step in supporting a
complete the network suitability successful system deployment. A system
analysis. design team should review current GIS and
• Platform architecture selection—CPT hardware system technology, review user
Design Platform tier is configured to requirements, and establish a system
represent the design solution. Identify architecture design based on user workflow
platform tier nicknames, select needs. A deployment schedule, as shown in
platforms, and identify platform rollover Figure 12.11, should be developed to
settings. identify overall implementation objectives.
• Software configuration—CPT Design
Software Configuration module is used Phased implementation strategies can
to assign workflow software to significantly reduce implementation risk.
supporting platform tier (software install) Computer technology continues to evolve at
and make workflow data source a remarkable pace. Integration standards
selection. are constantly changing with technology
• Enterprise design solution—Once and, at times, may not be ready to support
configured, the CPT Design tab immediate system deployment needs. New
completes the system architecture ideas are introduced into the market place
design analysis and provides the every day, and a relatively small number of
platform solution. these ideas develop into dependable long-
term product solutions. The following best
practices are recommended to support a
successful enterprise GIS implementation.

System deployment phases • Product upgrade deployments are
integrated into production roll-out
Pilot Phase schedule when ready.
• Enterprise GIS environments are
• Represent all critical hardware upgraded incrementally to meet
components planned for the final system operational requirements.
solution. • Functional and performance testing is
• Use proven low-risk technical solutions completed before production roll-out.
to support full implementation. • Configuration control for each upgrade
• Include test efforts to reduce uncertainty is critical for implementation success.
and implementation risk.
• Qualify hardware solutions for initial ArcGIS Enterprise Deployment
production phase.
Enterprise deployment of ArcGIS setups is
Initial Production Phase now possible starting with the ArcGIS 10.2
release. With enterprise deployment of
• Do not begin until final acceptance of ArcGIS products, GIS managers or system
pilot phase. administrators can efficiently plan for and
• Deploy initial production environment. control ArcGIS installations and updates.
• Use technical solutions qualified during Any enterprise deployment tool that
the pilot phase. supports Windows installation packages
• Demonstrate early success and payoff (MSI files) can be used to deploy ArcGIS
of the GIS solution. software setups. ArcGIS 10.2 Enterprise
• Validate organizational readiness and Deployment technical paper discusses the
support capabilities. enterprise deployment of ArcGIS software
• Validate initial training programs and setups using Microsoft Systems
user operations. Management Server (SMS), System Center
• Qualify advanced solutions for final Configuration Manager (SCCM), and Group
implementation. Policy.
Final Implementation Phase Virtual Desktop and Server

Technology
• Do not begin until final acceptance of
initial production phase.
• Plan a phased roll-out with reasonable
slack for resolving problems.
• Use technical solutions qualified during
previous phases.
• Prioritize roll-out timelines to support Figure 12.12 Virtual server deployments provide
early success. options for adaptive virtual desktop environments for
development, controlled virtual server environments
for testing and staging, and protected high available
Implementation strategies are virtual server environments for production.
accelerating with faster technology
evolution Virtual server technology is reducing the
cost of managing a rapidly changing IT
• Production upgrades are scheduled to environment. Figure 12.12 identifies a
enable required technology deployment strategy taking advantage of
advancement. virtual server technology. Many ESRI
development and testing operations are

currently supported in virtual desktop or The real need for more rapid adaptive
virtual server environments. Vendors are deployment schedules (to keep pace with
improving management and performance changing technology) which also must be
monitoring of virtual server environments, coupled with more stable production
and it is becoming more practical to manage deployments (reduced production downtime
and deploy production environments in and more rapid failure recover) drive the
virtual server deployments. A majority of need for virtual platform environments -
GIS operations are being deployed today virtualization is one solution that addresses
with virtual server environments. some real IT management needs.
Platform virtualization technology provides The potential disadvantages can be

IT managers with a way to abstract the managed by proper deployment strategies.
installed platform software environment The performance overhead for virtual
from the physical platform hardware. There desktop environments is much higher than
are two fundamental levels of virtualization, for server environments, and for this reason
one being virtual desktop environments virtual desktops are normally limited to
hosted within a physical platform operating software development environments where
system that interfaces with the physical performance and scalability is not a critical
platform hardware and the other being a factor.
virtual server environment hosted on a
hyper-visor layer that interfaces with the Server consolidation benefits can be
assigned physical platform hardware. In leveraged in a Staging environment, were
both cases, the virtual desktop or server multiple release candidates can undergo
contains its own dedicated operating system test and validation in preparation for
and software install separate from other production deployment. Several production
virtual systems deployed on the same release candidates can be testing in parallel
hardware. on the same physical server platform.
There are many recognized benefits with Production deployment can benefit from
virtual desktop/server deployments. deploying an existing virtual server install
(Staging configuration that has completed
• The benefits include faster provisioning final test and acceptance) to a higher
times, physical server platform capacity production physical server by
consolidation, fast recovery from system simply moving the Staging server release to
failures, simplified production delivery the production platform. If there is a
and recovery, and optimum production failure identified after
configuration control. All of these deployment; it is a simple process to move
benefits directly contribute to lower the production environment back to the
overall systems management costs and previous release. Deploying virtual server
a more stable operating environment. staging environments to a physical server
• The disadvantages include additional production environment is also a viable
software cost and some performance option - ensuring optimum performance and
overhead. There may also be functional scalability for the production environment.
limitations (limited access to hardware Virtual server migration software is available
graphic cards and performance to accomplish these provisioning tasks
monitoring software) which in many during live operations with no production
cases can be managed with the proper downtime.
deployment selections.

Virtual server performance impacts will integrate with existing systems that are
depend on the workflow environment, and constantly changing, and initial
can vary between 10 - 40 percent (and implementation can be painful.
sometimes higher) in the most efficient
virtual server configurations. Hardware Some of these ideas deliver on their
platform performance has improved over 80 promises, and in time they provide
percent within the past two years, which significant productivity advantages and
more than overcomes the virtual server reduce overall cost of administration. Soon
processing overhead. The new servers also a new idea comes along that performs
provide higher capacity (discussed in better at reduced cost, and organizations
chapter 9), which opens the door wider for must move on to new frontiers leaving
server consolidation benefits. legacy systems behind.
Virtual server deployments appear to be Selecting the right technology at the right
moving to mainstream IT production time will optimize business performance.
environment. The big question is no longer Introducing new technology before it is
whether it makes sense to deploy on virtual ready for prime time can reduce productivity
servers, but rather when and which software and increase implementation cost. Delaying
vendor solution will provide the highest too long can result in missed opportunities.
return on investment. Getting the timing right promotes success.
Technology Product Life Cycle

System Testing
Conducting proper testing at the right time
can contribute to implementation success.
Functional component and system
integration testing should be conducted for
new technology during prototype
development and before introduction into
production. The primary focus during this
testing is to make sure everything works.
Figure 12.13 Software technology life cycle
Performance targets established during the
represents the risk facing current GIS managers.
Buying too early can cost more, and buying too late initial system design can be evaluated
impacts user productivity. during early testing, paying close attention
to map display performance and layer
Figure 12.13 provides an overview of the complexity (see Chapter 3 Software
technology product life cycle, from initial Performance). This is an opportunity to
introduction of a new idea (product evaluate workflow functions and reduce
innovation) through end of life. Technology processing overhead.
is changing faster every year, and
managing technology change within a Enterprise system environments are
production environment is a challenge for becoming more complex. Testing should be
GIS managers and IT administrators. conducted in a production software
environment (same operating system,
We are seeing an increasing number of new service pacts, software architecture, etc).
ideas introduced into the marketplace, with Configuration challenges such as firewall
each idea promising improved user access, security, and high availability should
productivity and simplified system be configured and tested before
administration. These new ideas must deployment. Development and test

environments should be established to Test planning:
represent the complete software
configuration for each production release • Complete a risk analysis: Identify new
cycle. functionality that requires testing.
• Identify test objectives and establish
Functional Testing configuration control plan.
• Identify test hardware and software
configuration.
• Develop test procedures.
Test implementation:
• Identify implementation team and

establish implementation schedule.
• Order hardware and software and
publish installation plan.
• Conduct test plan and validate
functional acceptance.
• Collect test performance parameters
(CPU, memory, network traffic, etc.).
Test results and documentation:
• Document the results of the testing.

o Include specific hardware,
Figure 12.14 Formal functional testing should be
completed before any system upgrades are software, and network
introduced into a production environment. components that were tested.
o Include installation and test
Figure 12.14 identifies best practices for procedures that were followed,
planning and conducting functional system test anomalies, and final
testing. Conducting proper testing at the resolution.
right time can contribute to implementation o Complete test compliance matrix
success. Functional component and system identifying validation of functional
integration testing should be conducted for requirements.
new technology during prototype • Publish the test results and make
development and before introduction into available for reference during system
production. implementation.
A test plan should be developed with clearly Best practice:

defined test requirements, establish
configuration control (software versions, • Complete prototype integration testing
operating system environment), and provide before production deployment.
test procedures. Testing should be • Test in a production environment
completed before production deployment. (configuration control).
Testing should be conducted using the • Document functional requirements and
software versions and operating system that test procedures.
will be deployed in the production
environment.

Performance Testing The scientific method introduced with grade
school science fair projects provides some
System test is a free desktop application fundamental best practices that directly
from Esri that enables GIS administrators apply to system performance testing.
and users to quickly and easily create Performance testing should only be
functional and load testing procedures conducted to validate a hypothesis
within a GIS environment. (something you think you know). The
primary objective of a performance test is to
validate the hypothesis (confirm what you
know). The test is a success only if it proves
the hypothesis (testing does not teach you
what you don't know).
Initial performance testing often fails to

support the test hypothesis. With further
analysis and investigation, test bottlenecks
and/or improper assumptions are identified
that change the test results. Performance
testing is only successful if it validates the
test hypothesis. A scientific approach to
testing (Scientific Method) can help develop
a true interpretation of the technology.
Technology is changing with each service
pack release, and this requires an open
mind and willingness to continually change
what we believe to be true.
Figure 12.15 Performance testing should be
conducted throughout system design, deployment, System performance testing is best
and during production and should always follow the
scientific method. conducted during the initial production
deployment. During this phase, real users
Figure 12.15 identifies some best practices doing real workflows can generate a real
associated with system performance user environment. Critical system
testing. components should be monitored during the
initial deployment to identify processing
bottlenecks and resolve system conflicts.
Performance testing can be expensive and
Compliance with design performance
the results misleading. Normally initial
targets can be validated by observing
system deployments need to be tuned and
system throughput and utilization during
optimized to achieve final performance
initial user operations. Initial deployment
goals. Often system performance
acceptance should include validation that
bottlenecks are identified and resolved
user workflow performance goals are met.
during initial deployment. Early application
development focuses primarily on functional
requirements, and performance tuning is not
complete until the final release. Actual user
workflow environments are difficult to
simulate, and test environments seldom
replicate normal enterprise operations.

Systems Integration The capacity planning tool can be used by
project managers to establish performance
Management milestones and validate performance
targets are met throughout deployment.
Peak users can be identified for different
project milestones, and system throughput
and server utilization can be reported at
each milestone to demonstrate performance
goals are met.
Performance Monitoring
Figure 12.16 Project schedule should be developed to
identify implementation milestones and schedule Performance validation tools were
dependencies.
discussed in Chapter 3 (ArcGIS for Server
analyze and preview map optimization tools
Basic project management practices
and the mxdperfstat performance
promote implementation success. Project
monitoring tools). The system performance
teams should be established, individuals
terms discussed in Chapter 10, particularly
should be assigned specific responsibilities,
the relationship between throughput (peak
a task plan should be developed to support
users or peak transaction rates) and
implementation planning, a configuration
utilization (server CPU or network
control plan and change control process
bandwidth utilization), can identify if the
should be established, and an
deployed solution is performing within the
implementation schedule should be
initial project performance milestone.
published to support project deployment
milestones.
Monitoring live performance metrics can
provide excellent validation that the system
A system architecture design can provide
environment is designed to support peak
the framework for establishing an
throughput loads. The challenge is to collect
implementation plan. The implementation
appropriate throughput and utilization
plan should be developed after final
metrics that represent actual business
selection of the hardware vendor solution.
workflow loads (what are the current system
Figure 12.16 provides a typical system
loads).
deployment schedule. Specific decision
milestones should be included in the
The ArcGIS for Server statistics tab was a
schedule and each major task effort clearly
very useful tool for evaluating service usage
identified.
time on a live active ArcGIS Server
platform. This tool is no longer included in
An implementation project manager should
the new ArcGIS 10.1 Server Manager. The
be assigned to make sure all tasks are well-
ArcGIS 10.1 Help provides an example
defined, and every participant has a clear
script that can be used to derive map
understanding of his/her responsibilities. A
service statistics from the ArcGIS Server
clear set of acceptance criteria should be
logs. The FINE grain ArcGIS Server
developed for each implementation task and
statistics track which services are drawn
a formal acceptance process followed to
and how long the draws take. The example
ensure integration issues are identified and
script queries the logs and writes statistics
resolved at the earliest opportunity.
on map service activity during the sample
period. Results can be opened in Excel for
final review and analysis.

Department (Business) managers can often your service and help you make more
identify the percentage of peak user loads informed decisions about its use.
(throughput) generated on the system
during high volume events – these are the
same high volume events that drive staffing Performance Validation
and business planning needs. System
performance monitoring tools can measure
and report platform CPU utilization and
network traffic during these high volume
events. The problem is that business
management and IT administration staff
seldom share these metrics. These live
performance measures are the most
accurate resource for understanding
whether the existing system environment
has the capacity to meet expected peak
system loads. Figure 12.17 Monitoring performance compliance
throughout development (map publishing), initial
Several third party products are available for deployment, and in production can reduce
system tuning and Web load testing. The implementation risk.
ArcGIS Enterprise Systems: Performance
and Scalability presentation by our Figure 12.17 shows the importance of
Enterprise test team shares our system establishing performance milestones and
tuning and testing experience and best validating performance during
practices. The ArcGIS for Server implementation. Measuring display
Performance and Scalability - Optimizing rendering time when authoring a service is
GIS Services presentation provides several the first opportunity to validate performance.
tips on configuring GIS applications and Measuring deployed service rendering time
data sources for optimum performance. is another opportunity to validate
performance. The CPT test tab is designed
Additional ArcGIS for Server system to translate throughput and utilization
administration monitoring applications are measurements to workflow service times
shared on the ArcGIS Resource center. that can be compared directly to the initial
project workflow performance targets.
Monitoring progress in meeting performance
• Popular Extents. This application plots
milestones can reduce deployment risk and
extents requested from a map service
ensure project delivery success.
as graphics in a JavaScript application.
By extracting log records from the
ArcGIS Server Administrator API in this When performance issues are identified
manner, you can make more informed early in deployment, proper adjustments
decisions about which areas of your can be made before impacting production
maps are most widely used. workflow productivity (simpler map displays
• Services Dashboard. This application - less layers or generalize layers with large
displays statistics usage for a service, number of features, reduce number of
such as the number of service instances current batch jobs during peak system
in use, average response time, and the loads, evaluate preprocessing alternatives
total number of errors, warnings, and (map cache, generalized geodatabase
successes that are logged. The layers, etc). Chapter 3 identified a variety of
Services Dashboard can be used to ways to improve GIS display performance.
more closely monitor the current state of Identifying and resolving performance
issues before they become production level

performance problems will promote System tuning is a critical part of final
deployment success. system integration and deployment. Initial
system deployment is the first opportunity to
System Monitor is is provided through Esri begin performance tuning. Heavy batch
Professional Services for monitoring and processing efforts should be identified and
analyzing your enterprise GIS system. separated from interactive user workflows
System Monitor capabilities are shared on and supported through a separate batch
the System Monitor Demo site. process queue. System backups and heavy
processing workloads should be scheduled
during off-peak workflow periods.
System Tuning
System component performance metrics
should be monitored on a periodic basis
particularly during peak workflow periods to
identify performance bottlenecks and
address system deficiencies. Figure 12.18
provides an overview of the components
supporting an enterprise GIS environment.
Any component has the potential to
Figure 12.18 System performance tuning is a way to introduce a weak link in the overall system
get the most out of your system investment. The performance equation.
weakest system component (chain link) will limit
system performance. Make sure each component is
pulling its own weight.

Managing Technology Change
Figure 12.19 Managing technology change is the biggest challenge for any GIS Manager. Building a GIS is an
iterative process requiring planning, test, and evaluation through each and every annual business cycle.
Enterprise GIS operations require a combination of strategic planning and a continued

investment in technology. Technology is changing very rapidly, and organizations that fail to
manage this change fall behind in productivity and overall cost management. Managing
technology change is a major IT challenge. Figure 12.19 identifies a conceptual system
architecture planning and deployment strategy for technology change management.
Enterprise operations should include a Planning, test, and validation:

periodic cycle of coordinated production
system updates. • Planning activities should be established
on a periodic cycle.
• Planning, test, and technology validation • Coordinated to support the
should occur one periodic cycle ahead organization's operational and budget
of each production deployment. planning needs.
• Production deployment efforts should be • Strategic plans should be updated to
coordinated to support operational support a multiyear deployment strategy
technology needs. and published periodically (normally on
an annual cycle).

The planning and evaluation process should System deployment:
include:
• Operational system upgrades should be
• A requirements evaluation (strategic planned on a periodic cycle.
plan update). • Scheduled to implement validated
• Technology refresh (training and operational enhancements from the
research). planning and evaluation program.
• Requirements analysis (process and • System deployment phases should
requirements review). include initial implementation
• Test and evaluation (evaluate new (implementing changes in an
technology alternatives). operational test environment) to support
• Prototype validation (pilot test deployment authorization.
programs). • The program should also include
planned schedules for new technology
Best practice: Efforts should be procurement and deployment on a
scheduled to support the annual system periodic schedule (in some cases
deployment upgrade cycle. deployment upgrades can be
implemented on a monthly or quarterly
basis).
• All production system upgrades should
be planned and scheduled with full
support for ongoing operations.
Conclusion
Successful implementation depends on a good solid design, appropriate hardware and software
product selection, successful systems integration, and careful incremental evaluation during
installation. A phased approach to implementation reduces project risk and promotes success,
providing the opportunity for early success and flexibility to incorporate new technology at low
risk prior to final system delivery.
Guidelines are available to support a successful system design, even for large complex
systems. Final purchase decisions are influenced by both operational requirements and budget
limitations, introducing unique challenges for system design. Good leadership, qualified staff,
and proven standard practices support successful deployments.

SDSwiki 2017

Uploaded by

Copyright:

Available Formats

SDSwiki 2017

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

SDSwiki 2017

Uploaded by

Copyright:

Available Formats

System Design Strategies

An Esri ® Technical Reference Document • Fall 2017

An Esri ® Technical Reference Document • Fall 2017

Specific license terms for this content

1. System Design Process

This System Design Strategies wiki Software Performance

GIS technology is rapidly evolving to leverage

System Design Strategies Preface 0.1

System Design Strategies Preface 0.2

System Design Process 1.1

System Design Process 1.2

System Design Process 1.3

System Design Process 1.4

System Design Process 1.5

A primary goal behind developing capacity

System Design Process 1.6

Best Practice: Business decisions for

System Design Process 1.7

Figure 1.11 Response time includes display

System Design Process 1.8

System Design Process 1.9

Valid workflow Figure 1.14 Batch process productivity depends on

System Design Process 1.10

System Design Process 1.11

System Design Process 1.12

Figure 1.19 System Architecture Design

System Design Process 1.13

Figure 1.21 The CPT Design tab includes

System Design Process 1.14

System Design Process 1.15

System Design Process 1.16

System Design Process 1.17

GIS Software Technology 2.1

GIS Software Technology 2.2

GIS Software Technology 2.3

GIS Software Technology 2.4

Figure 2.5 Standard use cases supported by GIS

Figure 2.5 shows the standard user roles

GIS Software Technology 2.5

Figure 2.7 ArcGIS Curator role.

Figure 2.7 shows the ArcGIS Curator user

Figure 2.8 ArcGIS Analyst role.

GIS Software Technology 2.6

ArcGIS Desktop is changing

Figure 2.10 3D visualization is available in CityEngine,

GIS Software Technology 2.7

Figure 2.11 ArcGIS is evolving to consume vector tiles

Figure 2.11 shows the new vector tile format.

GIS Software Technology 2.8

Figure 2.13 AGD wkstn, AGD wkstn FSvc, and

Figure 2.13 shows the types of ArcGIS

GIS Software Technology 2.9

AGD Citrix FSvc workflows CPT desktop workflows

• ArcGIS Desktop installed on host server ArcGIS Desktop workflow performance

GIS Software Technology 2.10

Web app user roles Figure 2.16 ArcGIS Viewer role.

Figure 2.16 shows the ArcGIS Viewer user

Figure 2.15 shows the ArcGIS Contributor ArcGIS Enterprise licensing

Figure 2.17 shows the ArcGIS Enterprise