Cloud Foundations - Notes

PG Program in Cloud Computing
Introduction
What is cloud computing?
1. On-Demand availability of computing system resources, especially like storage and computing
power.
2. Option to use some else’s computer via the internet.
Use a thin computer (Laptop) to connect to a remote computer which is powerful. (Client - servers)
But any failure on the powerful computer will bring down the client computers.
Hence a pool of computers is created and hosted and shared on an on-demand basis. The shared
resource will be virtualized and delivered. Most of the maintenance of hardware is done by the cloud
provider.
Impacts:
AWS -> 400Billion

IBM - > 100Billion
AWS is less than 20yrs old while IBM is more than 50yrs old.
The market believes cloud computing has to offer more value and hence the growth is exponential and
future growth is expected to be high.
Azure added much value to Microsoft.

Successful apps today like Whatsapp, Salasforce, Snowflake etc were harnessing the power of cloud in
their startup phase to grow exponentially. So the impact is by far most valued to business growth as well
faster delivery.
Why such Impact?:
Division of labour! (Specialization) - Gives better efficiency since the expert in one field sticks to it and
invests time improving it. The IT infra handling part separated from the core businesses allowing them to
focus on the business value rather than the infra setup and maintenance.
Examples:Uber - Lets people hire - doesn't own a taxi

Facebook - Doesn’t create content.
But both these examples provide the service of making the provider and the customer meet.
© 2013 - 2021 Great Lakes E-Learning Services Pvt. Ltd. All rights reserved
Business case for Cloud Computing

Business Case:
Division of labour, allows to decrease cost and increase efficiency.

Allows your technical team to get to market quickly, with low cost, while allowing them to exclusively to
focus on the core business applications.
Different types of Computer/Cloud
SWOT Analysis
Strengths Weaknesses
Low upfront cost Managing spend
Low upfront time(time to market) Governance and control
No Maintenance/upgrades Data resides in remote
Low to no IT staff Might be a shared resource
Allow risk sharing (unused resource handling) Limited customization(any option outside the
cloud providers scope)
Mirroring, High availability and security Long run ROI unsure
Scalable and flexible Multiple cloud integration
Option to abandon infra without cost leverage Evolving internal systems and requirement for
change
CAPEX vs OPEX (CAPEX is always high &

Avoidable)
SWOT
Opportunities Threats
Cloud integration Compliance and Regulation
All enterprise system can be on the cloud Denial of service attacks - cybersecurity
Analytics built on the cloud One point failure - data loss
Exciting new technologies and paradigms Weak identity - access management
Account hijacking
Some instances where the cloud is secure (physical security). Other instances where local is secure
depends on the consumer.
Site Reliability Engineering
The developer keeps on improving the apps while sysAdmins tend to hold a state of the application
without change.
So while both work together the conflict is always inevitable.
If we make a developer work with the ops team to enable changes seamlessly and add automations for
deployment and maintenance.
Then the generalised term of ‘devops engineer’ arises. While google termed it as site reliability engineer.
Capacity Planning
The server needs for a specific application/business can be approximated and configured. But when the
utilization of the server is less than the max capacity and when the utilization of the server is reaching
above 100% (outage) then the capacity planning fails.
To avoid this cloud offer 2most important factors -> ‘Scale’ - increase capacity with the demand and
‘flexibility’ - reduce capacity when demand falls.
The customer is devoid of the all over capacity of the resources and only uses what is required for them.
In the meantime, cloud providers have a separate use for the available resources on the cloud which is
not utilized. It is provided as a low cost computation service to other customers.
How is cloud computing needed to be used?

When it should be opted for.
How it would help in the organisation.
Which professional uses it and how it is used.
Business problems
Businesses would have to consider a lot of points pros and cons before taking the decision to implement
cloud computing and the problems they face while using on-premise or data center are lists here:
a. Infrastructure as a service is required
b. Expertise of managing the data center
c. Business focus should not shift to handling the infra rather it should be within the business
priority actions.
d. Availability at uncertain loads
e. Provisional latency
f. Elasticity
g. Redundancy(Failure tolerance)
h. Infra refresh(life cycle to hardware)
i. Long term/Lock in
j. Confidentiality
k. Compliance
l. Audit trail
The classical enterprise structure and requirements
All the requirements to handle a complete IT infrastructure is being shown. The list reminds of both the
physical and virtual(software) requirements.
Why cloud?
Utilization factor when compared to Datacenter is better in Cloud (Elasticity) and time incurred in
deploying the hardware is reduced
History
From the 1950's the programming logic remained the same but the architecture of the infra has been
improving and helping us speed up the processing and delivery.
Current Era - The Fifth Inflection Point – Virtualization and Cloud computing
Definitions
Style of computing in which massively scalable IT related capabilities are provided “as a service” using
internet technologies to multiple “external customers”
Pool of abstracted, highly scalable, and managed compute infrastructure capable of hosting
end-customer applications and billed by consumption.
Myths of cloud computing

1. T
here’s one single “Cloud”
2. A
ll you need is a credit card
3. T
he cloud always saves you money
4. T
he cloud always reduces your workload
5. I ntegration
a. S
eamlessly blend cloud and your on-premises
b. N
ot suitable to blend public and private cloud
6. C
loud provider can guarantee security
7. U
sing virtualization and cloud computing is same
8. C
loud computing is only about technology
There are myths only but also would come true if implemented without much expertise or knowledge in
it.
Service delivery models
Different models of service delivery have been explained here.
SPIDERS
SaaS
PaaS
IaaS
bigData
Elastic
Resilient
Subscription
Prospective pizza example
Example for different service models with a pizza example.
Cloud providers comparison
Degree of abstraction - app view
Same example over and application service.
Actual Datacenter view

Check out google datacenter:
https://www.youtube.com/watch?v=XZmGGAbHqa0
Cloud Computing attributes
What powers up cloud computing and makes organisation, developers and administrators be inclined
towards it. The pointers mentioned in the above picture add those values to the cloud.
Cloud Offerings
Since the type of tools used inside each cloud has its own name here are the generalized names of the
tools that each cloud provides. For example, elastic infrastructure is the same technology but used as
EC2 instances in AWS , virtual machines in Azure and GCP.
Hosted Managed services
The actions items of classical enterprise are now packaged as a cloud provider and hence all those
services become managed services. The cloud customer just needs those services as plug and play and
pay only for those services.
Cloud Deployment models

Public:
A model where a service provider makes resources, such as applications and storage, available to the
general public over the internet.
The hosting environment is shared between many customers possibly reducing the costs for an
individual customer.
Leveraging economies of scale enables a dynamic use of resources, because workload peaks of some
customers occur during times of low workload of other customers
It is hosted and managed by a 3rd party from one or more data centers
Private:
Cloud computing properties are enabled in a company-internal datacenter
Alternatively, the private cloud may be hosted exclusively in the data center of an external provider, then
referred to as an outsourced private cloud.
Public Cloud providers also offer means to create an isolated portion of their cloud made accessible to
only one customer: a Virtual Private cloud which is the default behavior for many public cloud providers.
Community:
IT resources required by all collaborating partners are offered in a controlled environment accessible
only by the community of companies that generally trust each other.
Carving out a dedicated area exclusively for a company could be possible(private)
A similar model to that of Asp, completely managed by a 3rd party.
Hybrid:
Any combination of Public, private and Community
Eg any cloud along with any static in-house data center are integrated.
Applications can choose the right environment leveraging the best from each option.
Enabled in “Cloud bursting”
Interconnecting usually happens via VPN
Private connect – no connection to actual internet.
Subscription model
1. C
ustomers avoid large upfront capital expense
2. P
ay as an ongoing operational expense
3. E
asily and quickly scale up or down based on business demand and only pay for what is needed.
4. B
etter matches today’s financial drivers
5. N
o buying, it’s always ‘Renting’
Cost economics – Classical mode
Cost economics – Cloud model
Delta T calculation for triggering the ‘add provision’ or ‘remove provision’ needs to be set by the user with
respect to the application needs and demand monitoring.
Note: Delta T is the time taken for a new resource to get added into the infra and act as added capacity
to the infra. Without proper calculation of which the capacity graph will look like steep steps up and step
down.
Vertical scaling (Specialized) vs. Horizontal scaling (Commodity)

Large Operation expense Less Operating expense
Wasted/Idle resource Maximum resource utilization
Failure takes out a large chunk Failure takes out a small chunk
Expensive redundancy model Inexpensive redundancy
One shoe fitting all model Specific h/w for specific task
Too much co-existence Very less to no co-existence
Introduction to Virtualization
Option to utilize a resource to fullest. Divide and rule formula, rather than having one big machine.
a. Virtualization of the computing resources, including servers, network and storage, allows dynamic
flexibility.
b. Capacity can be more efficiently utilized.
c. Quickly add new servers without delay due to procurement or installation.
d. Easy to turn on or off virtual servers to handle scalability.
e. Physical connectivity is done up front and configuration is done in software at provisional time.
f. Networking equipment and storage is virtualized as well.
A typical application stack
The single OS bears all the underlying bare metal to the application level.
Virtualized Stack
Multiple OS on the bare metal and application runs within an OS on the same machine but on different
virtual machines
Characteristics of VM in Cloud
Business Perspective:
1. Option of more computing resources available on demand
2. Elimination of upfront monetary commitment
3. Ability to pay for computing resources on a short term basis as needed.
4. There are options to acquire compute resources by ‘auction’
5. A service centric approach with self-service and is self-managed
Technology Perspective:
1. Partitioning: In virtualization, many applications and operating systems are supported in a single
physical system by partitioning the available resources.
2. Isolation: Each virtual machine is isolated from its host physical system and other virtualized
machines. Because of the isolation, if one virtual-instance crashes, it doesn’t affect the other
virtual machines. In addition, data isn’t shared between one virtual container and another.
3. Encapsulation: A virtual machine can be represented as a single file, so you can identify it based
on the service it provides. In essence, the encapsulated process could be a business service.
4. Flexible: Should be able to configure and reconfigure to meet the growing and changing needs.
Virtualization Drawbacks
Amplified physical failures
Skills required to set up perfectly
Complex root cause analysis
Estimating the number of VMs per physical hardware
Performance reduces with added layers
Some application do not play well in a virtualized environment
Licensing costs
Next gen Virtualization

1. Dynamic swapping in virtualization is a need to avoid amplified failures
2. Virtualized machines come with fixed sizes inside the fixed machines. i.e. max cores
cannot exceed the hardware machines limits.
3. How to fix the number of VMs per bare metal machine?
Containers :- (LinuX Containers – LXC)
1. Application focused bundling
2. Common aspects(OS, library etc.) are shared
3. Smaller footprints
4. Quick load time
Containerization vs Virtualization
Container vs Virtualization
All apps need to be deployed on VM can have a choice of

the same OS image (good OS(good option for Iaas)
choice for PaaS)
Containers have less isolation VM allows for higher isolation

among them
Containers will usually take VM may take minutes

milliseconds to start
Use Container or virtualization?

Depends on the business requirement totally
Use cases where both options are combined is also a valid scenario.
PaaS overview
Abstracts the infra layer and enables us to monitor both infra and application layers. Helps in managing
the underlying layers.
Cloud service Taxonomy
The main models being the primary IaaS, PaaS and SaaS, any higher level service models can be
created out of these and above and combination goes on.
Price economics & Decision making

Profit = Revenue – Costs (Since everything is pay-as-you-go)
Cloud allows you to control and monitor the ops costs
Cost Based:
Cost of building product
Cost of maintaining the product
Providing support for the product
Other factors that may be specific to the product as well as the organization
Charging for the services offered and/or enhanced
Pass through costs can find a way in the product price
Value based:
Customer perception about the product
What values customer derives from the product
Longevity of the product
Recognizable name associated with luxury
And hence the cost in different regions will be different.
And hence you will need to analyze the cost aligned with multiple factors and decide/plan the strategies.
Data to Information xform velocity

The decision to make the cost optimized needs to be analyzed according to your requirements and then
the cost should be calculated.
Cloud provides managed service which handles these cost optimization
Similarly many other such data collection is managed by cloud and provided as a managed service like
monitoring and dashboards etc.
Challenges with Distributed computing

Heterogeneity (Apps, DB, Cache, Storage)
Fault handling
Consistency – Strict and Eventual
Global concurrency
Upgrades and maintenance
Local file systems
Application sessions – transient data (synchronization)
Designing application for the cloud

1. I mpact analysis on architecture
2. W
hich cloud provider
3. F
actoring defining the business process and integration
4. S
ecurity options
5. F
oundation, stateless, future etc.
6. M
icro service approach
Distribution of Control between Service Models

1. D
ecentralized Administration – each service model retains administrative control over its resources
2. Secure Distributed Collaboration – SLA maintenance even when Cloud provider faces some service
offline
3. C
redential Federation – decentralized single-sign-on mechanism
4. P
lacement of functionality – huge impact if not placed properly
5. F
ederated Data Collaboration – use on-prem and cloud else bring everything inside cloud
6. Loose coupling – different evolution cycle for the different services and hence loose coupling will help
in retaining the infra intact and still act with individuality.
Security of architecture with multiple providers.

Authenticator like OAuth, facebook, Gmail can be used for user login/data.
Shared responsibility model

Security “of the cloud” – with cloud provider
Security of your application “in the cloud” – you are responsible
Diversity of programming languages

There are many languages and the cloud is compatible with anything you want until you fix the
architecture in the cloud accordingly. Popularity index:
1. Python
2. Javascript
3. Java
Infrastructure automation
Infrastructure provisioning can be automated using machine-readable scripts, instead of performing
manual provisioning. This leads to considerable increase in speed of infrastructural operations, and
eliminates the human error.This is known an Infrastructure as Code (IaC)
Options:
a. Chef.io
b. Puppet.com
c. Ansible
d. CI using Jenkins
e. Cloud provider based automation tools. Each has their own.
CI CD models
The developer pushes their codes to version control (git)
The CI/CD automatically kicks in when the code is pushed to git and publishes the code in staging and
starts testing/pushes to production, according to the pipeline flow design. Can also be used with a
docker.
IoT
Physical objects to interact with digital and connected through the internet. Ex. Smart refrigerators, smart
AC, smart water heaters, etc.
Google Brillo:-
https://www.youtube.com/watch?v=2rPkbyyviGI
Glossary
● Cloud Computing: The delivery of on-demand computing resources — everything from

applications to data centres — over the internet on a pay-for-use basis.
● Latency: Time interval between the stimulation and response.
● Elasticity: Elasticity is the ability to grow or shrink infrastructure resources dynamically as
needed to adapt to workload changes in an autonomic manner.
● Redundancy: Provision or existence of more than one means or resources to perform an activity
or function.
● Compliance: A state of being in accordance with established guidelines or specifications, or the
process of becoming so
● Business Process Management: The process of analysing and improving business processes
to create a more efficient and effective organization.
● ETL (Extract, Transform Load): A process in data warehousing responsible for pulling data out
of the source systems and placing it into a data warehouse.
● Microservices: An architectural style that structures an application as a collection of loosely
coupled services, which implement business capabilities.
● Resilient: Strong and able to recover from damage quickly.
● SPIDERS: SaaS PaaS IaaS bigDa ta Elastic Resilient Subscription
● IaaS (Infrastructure as a Service): A standardized, highly automated offering, where compute
resources, complemented by storage and networking capabilities are owned and hosted by a
service provider and offered to customers on-demand.
● PaaS (Platform as a Service): A broad collection of application infrastructure (middleware)
services (including application platform, integration, business process management and database
services).
● SaaS (Software as a Service): A s oftware that is owned, delivered and managed remotely by
one or more providers.
● Resource Pooling: Resources pooled to serve multiple consumers using a multi-tenant model,
with different physical and virtual resources dynamically assigned and reassigned according to
consumer demand.
● Managed Service: An information technology (IT) task provided by a third-party contractor and
delivered to a customer.
● Public Cloud: Computing services offered by third-party providers over the public Internet,
making them available to anyone who wants to use or purchase them.
● Private Cloud: Computing services offered either over the Internet or a private internal network
and only to select users instead of the general public.
● Community Cloud: A shared cloud computing service environment that is targeted to a limited
set of organizations or employees.
● Hybrid Cloud: A hybrid cloud is an integrated cloud service utilising both private and public
clouds to perform distinct functions within the same organisation.
● Vertical Scalability: Addition of resources to a single system node, such as a single computer or
network station, which often results in additional CPUs or memory.
● Horizontal Scalability: Ability to connect multiple hardware or software entities, such as servers,
so that they work as a single logical unit.
● Encapsulation: It is the mechanism that binds together code and the data it manipulates.
● Containerization: an OS-level virtualization method used to deploy and run distributed
applications without launching an entire VM for each app.
● Session Affinity: Session affinity overrides the load-balancing algorithm by directing all requests
in a session to a specific application server.
● Stateless: Class that does not have any class level attribute or an object that is instantiated,
used and thrown away e.g. HTTP

Cloud Foundations - Notes

Uploaded by

Copyright:

Available Formats

Cloud Foundations - Notes

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

Cloud Foundations - Notes

Uploaded by

Copyright:

Available Formats

What are the different types of cloud computing?

What are the different types of cloud computing?

What are some of the strengths and weaknesses of cloud computing?

What are some of the strengths and weaknesses of cloud computing?

PG Program in Cloud Computing

AWS -> 400Billion

Azure added much value to Microsoft.

Why such Impact?:

Examples:Uber - Lets people hire - doesn't own a taxi

Business case for Cloud Computing

Division of labour, allows to decrease cost and increase efficiency.

Different types of Computer/Cloud

Low upfront cost Managing spend

Low upfront time(time to market) Governance and control

No Maintenance/upgrades Data resides in remote

Low to no IT staff Might be a shared resource

Mirroring, High availability and security Long run ROI unsure

Scalable and flexible Multiple cloud integration

CAPEX vs OPEX (CAPEX is always high &

Cloud integration Compliance and Regulation

Analytics built on the cloud One point failure - data loss

Exciting new technologies and paradigms Weak identity - access management

Site Reliability Engineering

How is cloud computing needed to be used?

The classical enterprise structure and requirements

Myths of cloud computing

Service delivery models

Different models of service delivery have been explained here.

Prospective pizza example

Example for different service models with a pizza example.

Cloud providers comparison

Degree of abstraction - app view

Same example over and application service.

Actual Datacenter view

Cloud Computing attributes

Hosted Managed services

Cloud Deployment models

Cloud computing properties are enabled in a company-internal datacenter

Carving out a dedicated area exclusively for a company could be possible(private)

A similar model to that of Asp, completely managed by a 3​rd​ party.

Any combination of Public, private and Community

Enabled in “Cloud bursting”

Interconnecting usually happens via VPN

Private connect – no connection to actual internet.

Cost economics – Classical mode

Cost economics – Cloud model

Vertical scaling (Specialized) ​vs.​ Horizontal scaling (Commodity)

Wasted/Idle resource Maximum resource utilization

Expensive redundancy model Inexpensive redundancy

Too much co-existence Very less to no co-existence

b. Capacity can be more efficiently utilized.

c. Quickly add new servers without delay due to procurement or installation.

d. Easy to turn on or off virtual servers to handle scalability.

f. Networking equipment and storage is virtualized as well.

A typical application stack

1. Option of more computing resources available on demand

2. Elimination of upfront monetary commitment

3. Ability to pay for computing resources on a short term basis as needed.

4. There are options to acquire compute resources by ‘auction’

A similar model to that of Asp, completely managed by a 3rd party.

Vertical scaling (Specialized) vs. Horizontal scaling (Commodity)

● Cloud Computing: The delivery of on-demand computing resources — everything from