Introduction to computing with distributed data processing(UDS23201J) :Unit 1- Distributed Systems

Unit 1: INTRODUCTION TO DISTRIBUTED SYSTEMS

1.1 WHAT IS DISTRIBUTED SYSTEM AND DISTRIBUTED COMPUTING?

The process of computation was started from working on a single processor. This uni-
processor computing can be termed as centralized computing. As the demand for the
increased processing capability grew high, multiprocessor systems came to existence. The
advent of multiprocessor systems, led to the development of distributed systems with high
degree of scalability and resource sharing. The modern day parallel computing is a subset of
distributed computing.

A distributed system is a collection of independent computers, interconnected via

a network, capable of collaborating on a task. Distributed computing is computing
performed in a distributed system.

Distributed computing is widely used due to advancements in machines and faster

and cheaper networks. In distributed systems, the entire network will be viewed as a
computer. The multiple systems connected to the network will appear as a single system to the
user. Thus the distributed systems hide the complexity of the underlying architecture to the
user.
The definition of distributed systems deals with two aspects that:
 Deals with hardware: The machines linked in a distributed system are
autonomous.
 Deals with software: A distributed system gives an impression to the users that
they are dealing with a single system.
Features of Distributed Systems:
 Communication is hidden from users
 Applications interact in uniform and consistent way
 High degree of scalability
 A distributed system is functionally equivalent to the systems of which it is
composed.
 Resource sharing is possible in distributed systems.
 Distributed systems act as fault tolerant systems
 Enhanced performance
Introduction to computing with distributed data processing(UDS23201J) :Unit 1- Distributed Systems

Issues in distributed systems

 Concurrency
 Distributed system function in a heterogeneous environment. So adaptability is
a major issue.
 Latency
 Memory considerations: The distributed systems work on both local and shared
memory.
 Synchronization issues
 Applications must need to adapt gracefully without affecting other parts of the
systems in case of failures.
 Since they are widespread, security is a major issue.
 Limits imposed on scalability
 They are less transparent.
 Knowledge about the dynamic network topology is a must.

QOS parameters
The distributed systems must offer the following QOS:
 Performance

 Reliability
 Availability
 Security
Differences between centralized and distributed systems

Centralized Systems Distributed Systems

In Centralized Systems, several jobs are In Distributed Systems, jobs are distributed
done on a particular central processing among several processors. The Processor are
unit(CPU) interconnected by a computer network
They have shared memory and shared They have no global state (i.e.) no shared
variables. memory and no shared variables.

Clocking is present. No global clock.

Introduction to computing with distributed data processing(UDS23201J) :Unit 1- Distributed Systems

Fig 1.1: Distributed and Centralized systems

1.2 SOME EXAMPLES OF DISTRIBUTED SYSTEMS
Today’s internet exists on the distributed systems. There are numerous applications of
distributed systems. Some of them are discussed below:
Web Search
 The task of a web search engine is to index the entire contents of the World Wide
Web.

 Distributed search is a search engine model in which the tasks of Web crawling,
indexing and query processing are distributed among multiple computers and
networks.
 The search engines were supported by a single supercomputer . But in recent years,
they have moved to a distributed model.
 Google search relies upon thousands of computers crawling the Web from multiple
locations all over the world.
 In Google's distributed search system, each computer involved in indexing crawls
and reviews a portion of the Web, taking a URL and following every link available
from it.
 The computer gathers the crawled results from the URLs and sends that
information back to a centralized server in compressed format.
 The centralized server then coordinates that information in a database, along with
information from other computers involved in indexing.
 When a user types a query into the search field, Google's domain name server (
DNS ) software relays the query to the most logical cluster of computers, based on
factors such as its proximity to the user or how busy it is.
 At the recipient cluster, the Web server software distributes the query to hundreds
or thousands of computers to search simultaneously.
 Hundreds of computers scan the database index to find all relevant records.
 The index server compiles the results, the document server pulls together the titles
and summaries and the page builder creates the search result pages.
Introduction to computing with distributed data processing(UDS23201J) :Unit 1- Distributed Systems

 The following features of Google search makes it act as distributed system:

 The physical infrastructure with very large numbers of networked computers .
 Highly distributed file system that supports very large files.
 Availability of structured distributed storage system for fast access to data.
 Distributed locking and agreement.
 Works on a programming model that supports the management of large
parallel and distributed computations.

Massively multiplayer online games (MMOGs)

 These games simulate real-life as much as possible.
 As such it is necessary to constantly evolve the game world using a set of laws.
 These laws are a complex set of rules that the game engine applies with every clock
tick.
 The virtual world consists not only of human players but also of all game elements
that are not living objects.
 These elements are immutable and include the area terrain, trees, mountains,
rivers,
etc.
 MMOGs must be able to handle a very large number of simultaneous users.
 As the information transferred between the players and the game server is large,
the bandwidth required to support a huge number of players if enormous.
 Very large virtual worlds require huge computational power to simulate the
existence of life (AI Algorithms).
 No single processor machine can handle the computational load required.
 These gaming applications work on both client server and distributed architectures.
 In distributed architectures, the large virtual world of the game is split into
different smaller areas and each area is to be handled by a separate physical
machine (server).
 Therefore both the bandwidth and computational load is spread out on many
machines, thus making the application distributed.
Financial Trading
 The financial trading is now moving to distributed systems.
 These systems require frequent modifications, in response to the communication.
 Processing the events in distributed systems, demands reliability.
 So distributed event based systems are used.
Introduction to computing with distributed data processing(UDS23201J) :Unit 1- Distributed Systems

 A series of event feeds are given by the financial institution.

 The events may be in different formats, employ different technologies etc.
 So proper adaptors are a must.
 Pattern detection is also very import aspect in financial trading.
 The Complex Event Processing (CEP) is an automatized way of composing
event together into logical, temporal or spatial patterns.

Fig 1.2: Financial Trading System

Introduction to computing with distributed data processing(UDS23201J) :Unit 1- Distributed Systems

1.3 TRENDS IN DISTRIBUTED SYSTEMS

Distributed systems are undergoing a period of significant change and this can be
traced back to a number of influential trends:
• the emergence of pervasive networking technology;
• the emergence of ubiquitous computing coupled with the desire to support
user mobility in distributed systems;
• the increasing demand for multimedia services;
• the view of distributed systems as a utility.

1.3.1 Pervasive networking and the modern Internet

Pervasive computing devices are completely connected and constantly available.

 The products that are connected to the pervasive network are easily available.
 The main goal of pervasive computing is to create an environment where the
connectivity of devices is embedded in such a way that the connectivity is
unobtrusive and always available.
 Internet can be seen as a large distributed system.

Fig 1.3: Internet

 Internet is actually a collection of numerous sub networks operated by companies
and other organizations and typically protected by firewalls.
 The role of a firewall is to protect an intranet by preventing unauthorized messages
from leaving or
entering.
Introduction to computing with distributed data processing(UDS23201J) :Unit 1- Distributed Systems

 Internet Service Providers (ISPs) are companies that provide broadband links and
other types of connection to individual users and small organizations, enabling
them to access services anywhere in the Internet as well as providing local services
such as email and web hosting.

 The intranets are linked together by backbones network link with a high
transmission capacity, employing satellite connections, fibre optic cables and other
high-bandwidth circuits.

1.3.2 Mobile and ubiquitous computing

 Internet is now build into numerous small devices from laptops to watches.

 These devices must have high degree of portability. Mobile computing supports
this.

 Mobile computing is the performances of computing tasks while the user

dynamically changing his geographic location.

 Ubiquitous computing (small computing ) means that all small computing

devices will eventually become so pervasive in everyday objects.

Differences between ubiquitous computing and mobile computing

Ubiquitous Computing Mobile Computing

They could connect tens/hundreds of They could connect a few devices for every
computing devices in every room/person, person, small enough to carry around –
becoming “invisible” and part of the devices connected to cellular networks or
environment – WANs, LANs, PANs – WLANs
networking in small spaces

They could connect even the non- mobile They are actually a subset of ubiquitous
devices and offer various forms of computing.
communication.

They could support all form of devices They support only conventional, discrete
that are connected to the internet from computers and devices.
laptops to watches.
Introduction to computing with distributed data processing(UDS23201J) :Unit 1- Distributed Systems

Fig 1.4: Mobile devices in an environment

 The user in the above given environment has access to three forms of wireless
connection:
 The laptop connecting to the host’s wireless LAN.
 Mobile (cellular) telephone connected to the Internet. The phone gives access
to the Web and other Internet services.
 Digital camera that communicates over a personal area wireless network.
 This scenario demands associations between devices are routinely created and
destroyed called as spontaneous interoperation. This interoperation must be fast
and convenient.

1.3.3 Distributed multimedia systems

 A distributed system supports the storage, transmission and presentation of discrete

media types.
 A distributed multimedia system should be able to perform the same functions for
continuous media types such as audio and video.
 It should be able to store and locate audio or video files, to transmit them across
thenetwork to support the presentation of the media types to the user and
optionally also to share the media types across a group of users.
 The processing of such media files includes handling of temporal dimensions and
integrity of media.
 The distributed multimedia computing allows a wide range of new multimedia
services and applications to be provided on the desktop.
Introduction to computing with distributed data processing(UDS23201J) :Unit 1- Distributed Systems

 This includes access to live or pre-recorded television broadcasts, access to film

libraries offering video-on-demand services, access to music libraries, the
provision of audio and video conferencing facilities and integrated telephony
features including IP telephony or related technologies such as Skype, a peer-to-
peer alternative to IP telephony .

 Webcasting is an application of distributed multimedia technology.

Webcasting is the ability to broadcast continuous media, typically audio or

video, over the Internet.

 Webcasting demands the following changes in the infrastructure:

 Supportfor an range of encoding and encryption formats.
 Support for range of mechanisms to ensure that the desired quality of service
canbe met.
 Adaptability to associated resource management strategies.
 Providing adaptation strategies to deal with the situation where QOS is
difficlut to achieve.

1.3.4 Distributed computing as a utility

 Distributed systems are seen as a utility like water and electricity.
 The resources are provided by appropriate service suppliers and effectively rented
by the end user.
 The services may be physical or logical services.
 Physical resources such as storage and processing can be made available to
networked computers through data centres.
 Operating system virtualization is a key enabling technology that
users may actually be provided with services by a virtual rather than a physical
node.

 Software services rental, services such as email and distributed calendars.

 Cloud computing is the child of resource sharing.

A cloud is defined as a set of Internet-based application, storage and computing

services sufficient to support most users’ needs, thus enabling them to largely
or totally dispense with local data storage and application softwarewith local
data storage and application software.
Introduction to computing with distributed data processing(UDS23201J) :Unit 1- Distributed Systems

 Clouds are generally implemented on cluster computers.

 A cluster computer is a set of interconnected computers that cooperate closely to
provide a single, integrated high performance computing capability.
1.4 RESOURCE SHARING
 An important goal of a distributed system is to effectively utilize the collective
resources of the system, namely, the memory and the processors of the individual
nodes.
 Users think in terms of shared resources such as a search engine without regard for
the server or servers that provide these.
 A search engine on the Web provides a facility to users throughout the world, users
who need never come into contact with one another directly.
 In computer-supported cooperative working (CSCW), a group of users who
cooperate directly share resources such as documents in a small, closed group.
 The pattern of sharing and the geographic distribution of particular users determine
what mechanisms the system must supply to coordinate users’ actions.
 The service manages a collection of related resources and presents their
functionality to users and applications.
 Resources in a distributed system are physically encapsulated within computers
and can only be accessed from other computers by means of communication.
 Sharing is done by a program that offers a communication interface enabling the
resource to be accessed and updated reliably and consistently.
 In a client server paradigm, server refers to a running program on a networked
computer that accepts requests from programs running on other computers to
perform a service and responds appropriately.
 The requesting processes are referred to as clients.
 A complete interaction between a client and a server, from the point when the
client sends its request to when it receives the server’s response, is called a remote
invocation.
1.5 CHALLENGES IN DISTRIBUTED SYSTEMS
a) Heterogeneity
Heterogeneity means the diversity of the distributed systems in terms of hardware,
software, platform, etc. Modern distributed systems will likely to be operating with
different:
 Hardware devices: computers, tablets, mobile phones, embedded devices, etc.
Introduction to computing with distributed data processing(UDS23201J) :Unit 1- Distributed Systems

 Operating System: Ms Windows, Linux, Mac, Unix, etc.

 Network: Local network, the Internet, wireless network, satellite links, etc.
 Programming languages: Java, C/C++, Python, PHP, etc.
 Different roles of software developers, designers, system managers

 Middleware: Middleware applies to a software layer that provides a

programming abstraction as well as masking the heterogeneity of the
underlying
networks, hardware, operating systems and programming languages. Eg:
CORBA, RMI.
 Heterogeneity in mobile code: Mobile code is used to refer to program code
that can be transferred from one computer to another and run at the destination.
Eg: Java applets.

Fig 1.5: Challenges in Distributed systems

b) Transparancy
Distributed systems designers must hide the complexity of the systems. Adding
abstraction layer is particularly useful in distributed systems. While users hit
search in google.com, they never notice that their query goes through a complex
process before google shows them a result. Some terms of transparency in
distributed systems are:

Transparency Description

Access Hide differences in data representation and how a resource is accessed

Location Hide where a resource is located

Introduction to computing with distributed data processing(UDS23201J) :Unit 1- Distributed Systems

Migration Hide that a resource may move to another location

Relocation Hide that a resource may be moved to another location while in use

Replication Hide that a resource may be copied in several places

Concurrency Hide that a resource may be shared by several competitive users

Failure Hide the failure and recovery of a resource

Persistence Hide whether a (software) resource is in memory or a disk

The access and location transparency is collectively referred as network

transparency.
c) Openness
If the well-defined interfaces for a system are published, it is easier for developers to
add new features or replace sub-systems in the future. Example: Twitter and Facebook
have API that allows developers to develop their software. The following are key
points in openness:
 key interfaces are published.
 uniform communication mechanism and published interfaces for access to
shared resources.
 Open distributed systems can be constructed from heterogeneous hardware and
software, possibly from different vendors.
d) Concurrency
Distributed Systems usually is multi-users environment. In order to maximize
concurrency, resource handling components should anticipate as they will be accessed
by competing users. Concurrency prevents the system to become unstable when users
compete to view or update data.

e) Security
Every system must consider strong security measurement. Distributed Systems
somehow deals with sensitive information; so secure mechanism must be in place. The
foolowing attacks are more common in distributed systems:
 Denial of service attacks: When the requested service is not available at the time
of request it is Denial of Service (DOS) attack. This attack is done by bombarding
the service with a large number of useless requests that the serious users are unable
to use
it.
Introduction to computing with distributed data processing(UDS23201J) :Unit 1- Distributed Systems

 Security of mobile code: Mobile code needs to be handled with care since they
are transmitted in an open environment.
f) Scalability
Distributed systems must be scalable as the number of user increases.

A system is said to be scalable if it can handle the addition of users and

resources without suffering a noticeable loss of performance or increase
in administrative complexity.

Scalability has 3 dimensions:

 Size: Number of users and resources to be processed. Problem associated with this
is overloading.
 Geography: Distance between users and resources. Problem associated is
communication reliability
 Administration: As the size of distributed systems increases, many of the system
needs to be controlled. Problem associated is administrative mess
Scalability often conflicts with small system performance. Claim of scalability in such
system is often abused. The following are the scalability challenges:
 Controlling the cost of physical resources
 Controlling the performance loss
 Preventing software resources running out
 Avoiding performance bottlenecks

g) Resilience to Failure (Fault tolerance)

Distributed Systems involves a lot of collaborating components (hardware, software,
communication). So there is a huge possibility of partial or total failure. The failures
are handled in series of steps:
 Detecting failures: Some failures like checksum can be detected.

 Masking failures: Some failures that have been detected can be hidden or
made less severe. Examples of hiding failures include retransmission of
messages and maintaining a redundant copy of same data.
 Tolerating failures: All the failures cannot be handled. Some failures must be
accepted by the user. Example of this is waiting for a video file to be streamed
in.
Introduction to computing with distributed data processing(UDS23201J) :Unit 1- Distributed Systems

 Recovery from failures: Recovery involves the design of software so that the
state
of
permanent data can be recovered or rolled back after a server has crashed.
 Redundancy: Services can be made to tolerate failures by the use of redundant
components. Examples for this includes: maintenance of two different paths
between same source and destination.
 Availability is also a major concern in the fault tolerance. The availabilityof a
system is a measure of the proportion of time that it is available for use. It is a
useful performance metric.

h) Quality of Service:
The distributed systems must confirm the following non functional requirements:
 Reliability:A reliable distributed system is designed to be as fault tolerant as
possible. Reliability is the quality of a measurement indicating the degree to
which the measure is consistent.
 Security: Security is the degree of resistance to, or protection from, harm. It
applies to any vulnerable and valuable asset, such as a person, dwelling,
community, nation, or organization. Distributed systems spread across wide
geographic locations. So security is a major concern.
 Adaptability:The frequent changing of configurations and resource
availability demands the distributed system to de highly adaptable.

1.6 WORLD WIDE WEB (WWW)

1.6.1 History and Development of WWW
 The World Wide Web (WWW) can be viewed as a huge distributed system
consisting of millions of clients and servers for accessing linked documents.
 Servers maintain collections of documents, while clients provide users an easy to-
use interface for presenting and accessing those documents.
 The Web started as a project at CERN, the European Particle Physics Laboratory
in Geneva, to let its large and geographically dispersed group of researchers
provide access to shared documents using a simple hypertext system.
 A document in a WWW could be anything that could be displayed on a user’s
computer terminal, such as personal notes, reports, figures, blueprints, drawings,
and so on.
 By linking documents to each other, it became easy to integrate documents from
different projects into a new document without the necessity for centralized
changes.
Introduction to computing with distributed data processing(UDS23201J) :Unit 1- Distributed Systems

 The Web gradually grew worldwide encompassing sites other than high energy
physics, but popularity really increased when graphical user interfaces became
available, notably Mosaic (Vetter et al., 1994).

 Mosaic provided an easy-to-use interface to present and access documents by

merely clicking the mouse.

 A document was fetched from a server, transferred to a client, and presented on the
screen.

 To a user, there was conceptually no difference between a document stored locally

or in another part of the world. This is transparent distribution.

 Since 1994, Web developments are primarily initiated and controlled by the World
Wide Web Consortium, which is a collaboration between CERN and M.I.T.

 This consortium is responsible for standardizing protocols, improving

interoperability, and further enhancing the capabilities of the Web.

 The webpages are portable and open.

1.6.2 Components of WWW

 HTML

 The Web documents are expressed by means of a special language called

HyperText Markup Language (HTML).

 HTML provides keywords to structure a document into different sections.

 One of its most powerful features is the ability to express parts of a document
in the form of a script.

 When a document is parsed, it is internally stored as a rooted tree, called a

parse tree, in which each node represents an element of that document.

 Each node is required to implement a standard interface containing methods for

accessing its content, returning references to parent and child nodes, and so on.
This standard representation is also known as the Document Object Model or
DOM or dynamic HTML.

 An alternative language that also matches the DOM is XML (Extensible

Markup Language).

 XML is used only to structure a document; it contains no keywords.

 XML can be used to define arbitrary structures. In other words, it provides the
means to define different document
types.
Introduction to computing with distributed data processing(UDS23201J) :Unit 1- Distributed Systems

 Uniform Resource Locators (URL)

The pages of a website can usually be accessed from a simple Uniform Resource
Locator (URL) otherwise called as web address.

 URL's are a way of identifying information on a server.

 A URL gives the protocol, the domain, the directory, and even the file.
 A URL consists of the following parts:
 protocol (such as http:// or ftp://)
 host name (the Web server's IP address or domain name)
 directory (i.e. folder)
 file name
There are two forms of URL as listed below:
 Absolute URL:
Absolute URL is a complete address of a resource on the web. This completed
address comprises of protocol used, server name, path name and file name.
Example: http:// www.abc.com / xyz /index.htm.
 http is the protocol.
 abc.com is the server name.
 index.htm is the file name.
The protocol part tells the web browser how to handle the file. Other protocols
also that can be used to create URL are: FTP, https, Gophe, mailto, news
 Relative URL
Relative URL is a partial address of a webpage. Unlike absolute URL, the
protocol and server part are omitted from relative URL. Relative URLs are used
for internal links i.e. to create links to file that are part of same website as the
WebPages on which you are placing the link.
Example:
To link an image on abc.com/xyz/internet_referemce_models, we can use the
relative URL which can take the form like /internet_technologies/internet-
osi_model.jpg.
Introduction to computing with distributed data processing(UDS23201J) :Unit 1- Distributed Systems

 Hyper Text Transfer Protocol (HTTP)

 HTTP is a communication protocol.
 It defines mechanism for communication between browser and the web server.
It is also called request and response protocol because the communication
between browser and server takes place in request and response pairs.
 It is a stateless protocol (i.e.) the history of the communication between server
and client is not stored in any form.
HTTP Request
 HTTP request comprises of lines which contains: Request line, Header Fields
and Message body.
 The first line i.e. the Request line specifies the request method i.e. Get or Post.
 The second line specifies the header which indicates the domain name of the
server from where index.htm is retrieved.
HTTP Response
 Like HTTP request, HTTP response also has certain structure.
 HTTP response contains: Status line, Headers and Message body.
 Publishing a resource:
 The methods for publishing resources on the Web are dependent upon the web
server implementation.
 The simplest method of publishing a resource on the Web is to place the
corresponding file in a directory that the web server can access.
 It is common for such concerns to be hidden from users when they generate
content.
 The database or file system on which the product pages are based is
transparent.
 Downloaded code:
 The designers of web services require some service-related code to run inside
the browser, at the user’s computer.
 Javascript is an example for downloaded code.
 A Javascript enhanced page can give the user immediate feedback on invalid
entries, instead of forcing the user to check the values at the server, which
would take much
longer.
Introduction to computing with distributed data processing(UDS23201J) :Unit 1- Distributed Systems

 AJAX (Asynchronous Javascript And XML) is used in cases where

synchronization is not a major concern.
 Applet is also a form of downloaded code.
 It is an application which the browser automatically downloads and runs
when it fetches a corresponding web page.

 Web services
 Web services allow exchange of information between applications on the web.
 Using web services, applications can easily interact with each other.
 The web services are offered using concept of Utility Computing.
Introduction to computing with distributed data processing(UDS23201J) :Unit 1- Distributed Systems

REVIEW QUESTIONS
PART-A
1. Define centralized computing.
The process of computation was started from working on a single processor. This uni-
processor computing can be termed as centralized computing.

2. Define distributed systems.

A distributed system is a collection of independent computers, interconnected via a
network, capable of collaborating on a task. Distributed computing is computing
performed in a distributed system.

3. List the features of Distributed Systems.

 Communication is hidden from users
 Applications interact in uniform and consistent way
 High degree of scalability
 A distributed system is functionally equivalent to the systems of which it is
composed.
 Resource sharing is possible in distributed systems.
 Distributed systems act as fault tolerant systems
 Enhanced performance
4. What are the issues in distributed systems?
 Concurrency
 Distributed system function in a heterogeneous environment. So adaptability is
a major issue.
 Latency
 Memory considerations: The distributed systems work on both local and shared
memory.
 Synchronization issues
 Applications must need to adapt gracefully without affecting other parts of the
systems in case of failures.
 Since they are widespread, security is a major issue.
Introduction to computing with distributed data processing(UDS23201J) :Unit 1- Distributed Systems

 Limits imposed on scalability

 They are less transparent.
 Knowledge about the dynamic network topology is a must.
5. Mention the QOS parameters of DS.
The distributed systems must offer the following QOS:
 Performance
 Reliability
 Availability
 Security

6. Give the differences between centralized and distributed systems

Centralized Systems Distributed Systems

In Centralized Systems, several jobs are In Distributed Systems, jobs are
done on a particular central processing distributed among several processors.
unit(CPU) The Processor are interconnected by a
computer network
They have shared memory and shared They have no global state (i.e.) no
variables. shared memory and no shared
variables.
Clocking is present. No global clock.

7. What is distributed search?

Distributed search is a search engine model in which the tasks of Web crawling,
indexing and query processing are distributed among multiple computers and
networks.

8. List the features of Google DS.

The physical infrastructure with very large numbers of networked computers .
Highly distributed file system that supports very large files.

Availability of structured distributed storage system for fast access to data.

Distributed locking and agreement.
Works on a programming model that supports the management of large parallel and
distributed computations.
Introduction to computing with distributed data processing(UDS23201J) :Unit 1- Distributed Systems

9. What is MMOG?
These games simulate real-life as much as possible. As such it is necessary to
constantly evolve the game world using a set of laws.These laws are a complex set of
rules that the game engine applies with every clock tick

10. What is CEP?

The Complex Event Processing (CEP) is an automatized way of composing event
together into logical, temporal or spatial patterns.

11. What is pervasive computing?

Pervasive computing devices are completely connected and constantly available.

12. Define ISP.

Internet Service Providers (ISPs) are companies that provide broadband links and
other types of connection to individual users and small organizations, enabling them to
access services anywhere in the Internet as well as providing local services such as
email and web hosting.

13. Define Ubiquitous computing .

Ubiquitous computing (small computing ) means that all small computing
devices will eventually become so pervasive in everyday objects.
14. Give the differences between ubiquitous computing and mobile computing

Ubiquitous Computing Mobile Computing

They could connect tens/hundreds of They could connect a few devices for every
computing devices in every room/person, person, small enough to carry around –
becoming “invisible” and part of the devices connected to cellular networks or
environment – WANs, LANs, PANs – WLANs
networking in small spaces
They could connect even the non- mobile They are actually a subset of ubiquitous
devices and offer various forms of computing.
communication.
They could support all form of devices They support only conventional, discrete
that are connected to the internet from computers and devices.
laptops to watches.

15. Define webcasting.

Webcasting is the ability to broadcast continuous media, typically audio or video, over
the Internet.
Introduction to computing with distributed data processing(UDS23201J) :Unit 1- Distributed Systems

16. List the requirements of web casting.

Support for an range of encoding and encryption formats.
Support for range of mechanisms to ensure that the desired quality of service can be
met.
Adaptability to associated resource management strategies.
Providing adaptation strategies to deal with the situation where QOS is difficlut to
achieve.
17. Define cloud.
A cloud is defined as a set of Internet-based application, storage and computing
services sufficient to support most users’ needs, thus enabling them to largely or
totally dispense with local data storage and application software with local data
storage and application software.
18. What is CSCW?
In computer-supported cooperative working (CSCW), a group of users who cooperate
directly share resources such as documents in a small, closed group.
19. Define remote invocation.
A complDEete interaction between a client and a server, from the point when the
client sends its request to when it receives the server’s response, is called a remote
invocation.
20. List the challenges in DS.
 Heterogeneity
 Transparency
 Openness
 Concurrency
 Security
 Scalability
 Fault tolerance
 Providing QOS
21. What are the dimensions of scalability?
 Size: Number of users and resources to be processed. Problem associated with this
is overloading.
Introduction to computing with distributed data processing(UDS23201J) :Unit 1- Distributed Systems

 Geography: Distance between users and resources. Problem associated is

communication reliability
 Administration: As the size of distributed systems increases, many of the system
needs to be controlled. Problem associated is administrative mess
22. List the challenges in scalability.
The following are the scalability challenges:
 Controlling the cost of physical resources
 Controlling the performance loss
 Preventing software resources running out
 Avoiding performance bottlenecks

23. Give the features in fault tolerance.

Detecting failures: Some failures like checksum can be detected.
Masking failures: Some failures that have been detected can be hidden or made less
severe. Examples of hiding failures include retransmission of messages and
maintaining a redundant copy of same data.
Tolerating failures: All the failures cannot be handled. Some failures must be accepted
by the user. Example of this is waiting for a video file to be streamed in.
Recovery from failures: Recovery involves the design of software so that the state of
permanent data can be recovered or rolled back after a server has crashed.
Redundancy: Services can be made to tolerate failures by the use of redundant
components. Examples for this includes: maintenance of two different paths between
same source and destination.

24. What is parse tree?

When a document is parsed, it is internally stored as a rooted tree, called a parse tree,
in which each node represents an element of that document.

25. Define URL.

The pages of a website can usually be accessed from a simple Uniform Resource
Locator (URL) otherwise called as web address.
Introduction to computing with distributed data processing(UDS23201J) :Unit 1- Distributed Systems

PART – B
1. What are the issues in distributed systems?
2. Explain web search as a distributed system.
3. Brief about MMOGs.
4. Describe financial trading.
5. Explain the trends in distributed systems.
6. Elucidate the resource sharing nature of distributed system.
7. Brief about the challenges in distributed system.
8. Describe WWW.
9. Explain HTML.
 COMMUNICATION IN DISTRIBUTED
2
SYSTEM

2.1 COMMUNICATION IN DISTRIBUTED SYSTEMS

The most important difference between a distributed system and a uniprocessor system
is the interprocess communication.

Interprocess communication (IPC) is a set of programming interfaces that allows

a programmer to coordinate activities among different program processes that can
run concurrently in an operating system.

 In a uniprocessor system, interprocess communication assumes the existence of

shared memory whereas in a distributed system, there‟s no shared memory, so the
entire nature of interprocess communication must be completely reframed from
scratch.
 All communication in distributed system is based on message passing.
 The processes run on different machines and they exchange information through
message passing.
 Successful distributed systems depend on communication models that hide or
simplify message passing.

2.2 SYSTEM MODELS

System models describe common properties and design choices for distributed system
in a single descriptive model. The system models of distributed systems are classified into the
following

types:
 Distrubuted 2.2
Systems
 Physical models: It is the most explicit way in which to describe a system in
terms of hardware composition.
 Architectural models: They describe a system in terms of the computational and
communication tasks performed by its computational elements.
 Fundamental models: They examine individual aspects of a distributed system.
They are again classified based on some parameters as follows:
 Interaction models: This deals with the structure and sequencing of the
communication between the elements of the system
 Failure models: This deals with the ways in which a system may fail to
operate correctly
 Security models: This deals with the security measures implemented in the
system against attempts to interfere with its correct operation or to steal its
data.

2.2.1 Physical Models

A physical model is a representation of the underlying hardware elements of a

distributed system hides the details of the computer and networking
technologies employed.

There are many physical models available from the primitive baseline model to the
complex models that could handle cloud environments.
 Baseline physical model:
 This is a primitive model that describes the hardware or software components
located at networked computers.
 The communication and coordination of their activities is done by passing
messages.

 This is a minimal physical model of a distributed system.

 This model could be extended to a set of computer nodes interconnected by a

computer network for the required passing of messages.

 This model helps us to categorize the three generations of distributed systems.

 Early distributed systems:

 The period is in the late 1970s and early 1980s.

 This was developed after the emergence LAN.

 Distrubuted 2.3
Systems
 These systems could support 10 to 100 nodes interconnected by a LAN.

 It has very limited internet connectivity and can support a small range of
services such as shared local printers and file servers.

 Individual systems were homogeneous, with poor quality of service.

 Internet-scale distributed systems:

 The period is from the 1990s after the growth of internet.

 The physical set up of distributed system is described as an extensible set of

nodes interconnected by a network of networks (the Internet).

 They incorporate large numbers of nodes and provide distributed system

services.

 The systems were heterogeneous which could adopt open standards.

 They had good QOS.

 Contemporary distributed systems:

 The nodes were desktop computers and therefore relatively static, discrete and
independent.

 Mobile computing has led to physical models with movable nodes. Service
discovery is of primary concern in these systems.

 The cloud computing and cluster architectures hassled to pools of nodes that
collectively provide a given service. This resulted in enormous number of
systems given the service.

 Distributed systems of systems:

The ultra large scale (ULS) distributed systems is defined as a complex system
consisting of a series of subsystems that are systems in their own right and that come
together to perform a particular task or tasks. Some important characteristics of these
systems include their
 very large size
 global geographical distribution

 operational and managerial independence of their member systems.

The main function of these systems arises from the interoperability between their
components.
 Distrubuted 2.4
Systems
Example: Flood monitoring systems.

Parameter Early Systems Internet Scale ULS System

Scale Small Large Ultra Large
Heterogeneity Homogeneous Platform independent Supports many types
software and technologies. of architectures
Openness Systems are mostly Many open standards are The already existing
closed available standards are not able
to cope with complex
systems.
QOS Poor Significant QOS could still be
improved

2.2.2 Architectural Models

The architecture abstracts the functions of the individual components of the

distributed system.

This describes the components and their interrelationships to ensure the system is
reliable, manageable, adaptable and cost-effective. The different models are evaluated based
on the following criteria:
 architectural elements
 architectural patterns
 middleware platforms

2.2.2.1 Architectural elements

The following must be decided to build a distributed system:
 Communicating entities (objects, components and web services);
 Communication paradigms (inter process communication, remote invocation
and indirect communication);
 Roles responsibilities and placement
Communicating entities
The components that are communicating and how those entities communicate together
are dealt here. The following are some of the problem oriented
entities:
 Distrubuted 2.5
Systems
 Objects: They are used to implement object oriented approaches in distributed
systems. Objects are accessed through interfaces.
 Components: Components resemble objects and they offer problem-oriented
abstractions for building distributed systems. They are also accessed through
interfaces. The key difference between component and object is that a components
holds all the assumptions specified to other components and their interfaces in the
system. Components and objects are used to develop tightly coupled applications.
 Web services: Web services are closely related to objects and components. Web
services are integrated into the World Wide Web. They are partially defined by the
web-based technologies they adopt. Web services are generally viewed as
complete services that can be combined to achieve value-added services across
organizational boundaries.

Object Components Web services

Object oriented solution Problem oriented solution Problem oriented solution
The dependencies and The dependencies and They are integrated into the
assumptions of the interfaces assumptions of the all WWW.
holds only on the objects that interfaces holds on all the
act upon them. components in the system.
Tightly coupled services Tightly coupled services Complete service that could
also be made as a value added
service

Communication paradigms
There are three major modes of communication in distributed systems:
 Interprocess communication:
This is a low-level support for communication between processes in distributed
systems, including message-passing primitives. They have direct access to the API
offered by Internet protocols and support multicast communication.
 Remote invocation:
It is used in a distributed system and it is the calling of a remote operation,
procedure or method.
a) Request-reply protocols:
This is a message exchange pattern in which a requestor sends a request message
to a replier system which receives and processes the request, ultimately returning a
message in response. This is a simple, but powerful messaging pattern which
allows two applications to have a two-way conversation with one another over a
channel. This pattern is especially common in client–server architectures.
 Distrubuted 2.6
Systems
b) Remote procedure calls: Remote Procedure Call (RPC) is a protocol that one
program can use to request a service from a program located in another computer
in a network without having to understand network details.
c) Remote method invocation: RMI (Remote Method Invocation) is a way that a
programmer can write object-oriented programming in which objects on different
computers can interact in a distributed network. This has the following two key
features:
 Space uncoupling: Senders do not need to know who they are sending to
 Time uncoupling: Senders and receivers do not need to exist at the same time
 Indirect communication
Key techniques for indirect communication include:
a) Group communication: Group communication is concerned with the delivery
of messages to a set of recipients and hence is a multiparty communication
paradigm supporting one-to-many communication. A group identifier uniquely
identifies the group. The recipients join the group and receive the messages.
Senders send messages to the group based on the group identifier and hence do
not need to know the recipients of the message.
b) Publish-subscribe systems: This is a messaging pattern where senders
of messages, called publishers, do not program the messages to be sent directly
to specific receivers, called subscribers. Instead, published messages are
characterized into classes, without knowledge of what, subscribers there may
be. Similarly, subscribers express interest in one or more classes, and only
receive messages that are of interest, without knowledge of what, if any,
publishers there are. They offer one-to-many style of communication.
c) Message queues: They offer a point-to-point service whereby producer
processes can send messages to a specified queue and consumer processes can
receive messages from the queue or be notified of the arrival of new messages
in the queue. Queues therefore offer an indirection between the producer and
consumer processes.
d) Tuple spaces: The processes can place arbitrary items of structured data,
called tuples, in a persistent tuple space and other processes can either read or
remove such tuples from the tuple space by specifying patterns of interest. This
style of programming is known as generative communication.
e) Distributed shared memory: In computer architecture, distributed shared
memory (DSM) is a form of memory architecture where the (physically
separate) memories can be addressed as one (logically shared) address space.
Here, the term shared does not mean that there is a single centralized memory
but shared essentially means that the address space is shared (same physical
address on two processors refers to the same location in memory
 Distrubuted 2.7
Systems
Roles and responsibilities
There are two types of architectures based on roles and responsibilities:

1) Client-server:
 The system is structured as a set of processes, called servers, that offer services
to the users, called clients.
 The client-server model is usually based on a simple request/reply protocol,
implemented with send/receive primitives or using remote procedure calls
(RPC) or remote method invocation (RMI).
 The client sends a request (invocation) message to the server asking for some
service; - the server does the work and returns a result (e.g. the data requested
or an error code if the work could not be performed. )
 A server can itself request services from other servers; thus, in
relation, the server itself acts like a client. this new
 Some problems with client-server are centralization of service resulti
scaling, limited capacity of server bandwidth of network connecting t ng in poor
he server.

Figure 2.1: Client server style

2) Peer-peer
 All processes (objects) play similar role.
 Processes (objects) interact without particular distinction between clients and
servers.
 The pattern of communication depends on the particular application.
 A large number of data objects are shared; any individual computer holds only
a small part of the application database.
 Distrubuted 2.8
2.8 Communication in Distrubuted Sytem Systems
 Processing and communication loads for access to objec ts are distributed
across many computers and access links.
 This is the most general and flexible model.

Figure 2.2: Peer-peer style

 Peer-to-Peer model distributes shared resources widely share computing and
communication loads.
 Problems with peer-to-peer are high complexity due to cleverly place
individual objects and there are large number of replicas.

Placement
Mapping of objects or services to the underlying physical distributed infrastructure is
considered here. The following points must be taken care while placing the objects: reliability,
communication pattern, load, QOS etc. The following are the placement strategies:
 Mapping of services to multiple servers
 The servers may partition the set of objects on which the service is based and
distribute those objects between themselves, or they may maintain replicated
copies of them on several hosts.
 The Web provides a common example of partitioned data in which each web
server manages its own set of resources. A user can employ a browser to access
a resource at any one of the servers.
 Caching
 A cache is a local store of recently used data objects that is closer to one client
or a particular set of clients than the objects themselves.
 When a new object is received from a server it is added to the local cache store,
replacing some existing objects if
necessary.
 Distrubuted 2.9
Systems
 When an object is needed by a client process, the caching service first checks
the cache and supplies the object from there if an up-to-date copy is available.
If not, an up-to-date copy is fetched.
 Caches may be co-located with each client or they may be located in a proxy
server that can be shared by several clients.
 Mobile code:
 Applets are a well-known and widely used example of mobile code.
 The user running a browser selects a link to an applet whose code is stored on a
web server; the code is downloaded to the browser and runs.
 They provide interactive response.

 Mobile agents:
 A mobile agent is a running program (including both code and data) that
travels from one computer to another in a network carrying out a task on
someone‟s behalf, such as collecting information, and eventually returning
with the results.
 A mobile agent is a complete program, code + data, that can work (relatively)
independently.
 The mobile agent can invoke local resources/data.
 A mobile agent may make many invocations to local resources at each site it
visit.
 Typical tasks of mobile agent includes: collect information , install/maintain
software on computers, compare prices from various vendors bay visiting their
sites etc.
 Mobile agents (like mobile code) are a potential security threat to the resources
in computers that they visit.

2.2.2.2 Architectural patterns

Architectural patterns are reusable solution to a commonly occurring problem
in software architecture within a given context. They are not complete solutions but rather
offer partial insights. The following are some of the common architectural patterns:

 Layering:
 In a layered approach, a complex system is partitioned into a number of
layers, with a given layer making use of the services offered by the layer
below.
 Distrubuted 2.1
Systems 0

Figure 2.3: Layered Architecture

 A given layer is unaware of the implementation details, of other layers
beneath them.
 In terms of distributed systems, this equates to a vertical organization of
services into service layers.
 A distributed service can be provided by one or more server processes,
interacting with each other and with client processes in order to maintain a
consistent system-wide view of the service‟s resources.

 Tiered architecture
 Tiering is a technique to organize functionality of a given layer and place
this functionality into appropriate servers and, as a secondary
consideration, on to physical nodes.
 The two tiered architecture refers to client/server architectures in which the
user interface (presentation layer) runs on the client and the database
(data layer) is stored on the server. The actual application logic can run on
either the client or the
server.
 Distrubuted 2.1
Systems 1

Fig 2.4: Two and three tiered architectures

 The three tiered architecture has:
a) Presentation tier: This is the topmost level of the application. is
concerned with handling user interaction and updating the view of the
application as presented to the user;
b) Application tier (business logic, logic tier, or middle tier): The
logical tier is pulled out from the presentation tier and, as its own layer,
it controls an application‟s functionality by performing detailed
processing.
c) Data tier: The data tier includes the data persistence mechanisms
(database servers, file shares, etc.) and the data access layer that
encapsulates the persistence mechanisms and exposes the data.
 Thin clients
 In a thin-client model, all of the application processing and data
management is carried out on the server.
 The client is simply responsible for running the responsible for running the
presentation software.
 This is used when legacy systems are migrated to client server
architectures.
 Distrubuted 2.1
Systems 2
 The legacy system acts as a server in its own right with a graphical
interface implemented on a client.
 The major disadvantages of thin clients is that it places a heavy processing
load on both the server and the network and the delay due to operating
system latencies.
 The virtual network computing(VNC) has emerged to overcome the
disadvantages og thin clients.
 VNC is a type of remote-control software that makes it possible to control
another computer over a network connection.
 Keystrokes and mouseclicks are transmitted from one computer to another,
allowing technical support staff to manage a desktop, server, or other
networked device without being in the same physical location.
 Since all the application data and code is stored by a file server, the users
may migrate from one network computer to another. So VNC is of big use
in distributed systems.
 Other patterns
a) Proxy:
 This facilitates location transparency in remote procedure calls or remote
method invocation.
 A proxy is created in the local address space to represent the remote object
with same interface as the remote object.
 The programmer makes calls on this proxy object and he need not be aware
of the distributed nature of the interaction.
b) Brokerage:
 It is used to bring interoperability in potentially complex distributed
infrastructures.
 The service broker is meant to be a registry of services, and stores
information about what services are available and who may use them.

Fig 2.5: Web service with brokers

 Distrubuted 2.1
Systems 3
c) Reflection:
 The Reflection architectural pattern provides a mechanism for changing
structure and behavior of software systems dynamically.
 It supports the modification of fundamental aspects, such as type structures
and function call mechanisms.
 In this pattern, an application is split into two parts:
1) Meta Level: This provides information about selected system
properties and makes the software self-aware.
2) Base level: This includes the application logic. Its implementation
builds on the meta level. Changes to information kept in the meta level
affect subsequent base-level behavior.

2 2.2.3 Associated middleware solutions

Middleware is a general term for software that serves to "glue together" separate,
often complex and already existing, programs.

Middleware provides a higher-level programming abstraction for the development of

distributed systems. Middleware makes it easier for software developers to perform
communication and input/output, so they can focus on the specific purpose of their
application. Middleware is the software that connects software components or enterprise
applications and it lies between the operating system and the applications on each side of a
distributed computer network. Middleware includes Web servers, application servers, content
management systems, and similar tools that support application development and delivery.

Categories of middleware
The following are some of the categories of middleware:
 Distributed Objects
The term distributed objects usually refers to software modules that are designed to
work together, but reside either in multiple computers connected via a network or
in different processes inside the same computer. One object sends a message to
another object in a remote machine or process to perform some task. The results
are sent back to the calling object.
 Distributed Components
A component is a reusable program building block that can be combined with
other components in the same or other computers in a distributed network to form
an application. Examples: a single button in a graphical user interface, a small
interest calculator, an interface to a database manager. Components can be
 Distrubuted 2.1
Systems 4
deployed on different servers in a network and communicate with each other for
needed services. A component runs within a context called a container .
Examples: pages on a Web site, Web browsers, and word processors.
 Publish subscriber model
Publish–subscribe is a messaging pattern where senders of messages, called
publishers, do not program the messages to be sent directly to specific receivers,
called subscribers. Instead, published messages are characterized into classes,
without knowledge of what, if any, subscribers there may be. Similarly,
subscribers express interest in one or more classes, and only receive messages that
are of interest, without knowledge of what publishers are there.
 Message Queues
Message queues provide an asynchronous communications protocol, meaning that
the sender and receiver of the message do not need to interact with the message
queue at the same time. Messages placed onto the queue are stored until the
recipient retrieves them. Message queues have implicit or explicit limits on the size
of data that may be transmitted in a single message and the number of messages
that may remain outstanding on the queue.
 Web services
A web service is a collection of open protocols and standards used for exchanging
data between applications or systems. Software applications written in various
programming languages and running on various platforms can use web services to
exchange data over computer networks like the Internet in a manner similar to
inter-process communication on a single computer. This interoperability (e.g.,
between Java and Python, or Windows and Linux applications) is due to the use of
open standards.
 Peer to peer
Peer-to-peer (P2P) computing or networking is a distributed application
architecture that partitions tasks or work load between peers. Peers are equally
privileged, equipotent participants in the application. They are said to form a peer-
to-peer network of nodes.
Advantages of middleware
 Real time information access among systems.
 Streamlines business processes and helps raise organizational efficiency.
 Maintains information integrity across multiple systems.
 It covers a wide range of software systems, including distributed Objects and
components, message-oriented communication, and mobile application support.
 Middleware is anything that helps developers create networked applications.
 Distrubuted 2.1
Systems 5
Disadvantages of middleware
 Prohibitively high development costs.
 There are only few people with experience in the market place to develop and use
a middleware.

 There are only few satisfying standards.

 The tools are not good enough.
 Too many platforms to be covered.
 Middleware often threatens the real-time performance of a system.
 Middleware products are not very mature.

2.2.3 Fundamental Models

Fundamental Models are concerned with a formal description of the properties

that are common in all of the architectural models.

Models addressing time synchronization, message delays, failures, security issues are
addressed as:

 Interaction Model – deals with performance and the difficulty of setting of

time limits in a distributed system.

 Failure Model – specification of the faults that can be exhibited by processes

 Secure Model – discusses possible threats to processes and communication

channels.

The purpose of the fundamental model is to:

 make explicit all the relevant assumptions about the systems we are modelling.

 make generalizations (general purpose algorithms) concerning what is possible

or impossible with given assumptions.

2.2.3.1 Interaction model

 In this model the computations occurs within processes.

 The processes interact by passing messages to achieve communication

(information flow) and coordination (synchronization and ordering of activities)
between

processes.
 Distrubuted 2.1
Systems 6
 The two significant factors affecting interacting processes in a distributed system
are:

1) Communication performance
2) Global notion of time.

Communication Performance
 The communication channel in can be implemented in a variety of ways in
distributed systems: Streams or through simple message passing over a network.
 The performance characteristics of a network are:
a) Latency: A delay between the start of a message‟s transmission from one
process to the beginning of reception by another.
b) Bandwidth: The total amount of information that can be transmitted over in a
given time. The communication channels using the same network, have to
share the available bandwidth.
c) Jitter: The variation in the time taken to deliver a series of messages. It is very
relevant to multimedia data.

Global Notion of time

 Each computer in a distributed system has its own internal clock, which can be
used by local processes to obtain the value of the current time.
 Any two processes running on different computers can associate timestamp with
their events.
 However, even if two processes read their clocks at the same time, their local
clocks may supply different time because the computer clock drifts from perfect
time and their drift rates differ from one another.
 Even if the clocks on all the computers in a distributed system are set to the same
time initially, their clocks would eventually vary quite significantly unless
corrections are applied.
 There are several techniques to correct time on computer clocks.
 One technique is to use radio receivers to get readings from GPS (Global
Positioning System) with accuracy about 1 microsecond.

Variants of Interaction model

 In a DS it is hard to set time limits on the time taken for process execution,
message delivery or clock drift. There are two models based on time stamp:
 Distrubuted 2.1
Systems 7
 Synchronous DS
 This is practically hard to achieve in real life.
 In synchronous DS the time taken to execute a step of a process has
known lower and upper bounds.
 Each message transmitted over a channel is received within a known
bounded time.
 Each process has a local clock whose drift rate from real time has known
bound.
 Asynchronous DS
 There are no bounds on: process execution speeds, message transmission
delays and clock drift rates.

Event Modeling
 Event ordering is of major concern in DS.
 The concept of one event happening before another in a distributed system is
examined, and is shown to define a partial ordering of the events.
 The execution of a system can be described in terms of events and their ordering
despite the lack of accurate clocks.
 Consider a mailing list with users X, Y, Z, and A.
 Due to independent delivery in message delivery, message may be delivered in
different order.
 If messages m1, m2, m3 carry their time t1, t2, t3, then they can be displayed to
users accordingly to their time ordering.
 The mail box is:

Item From Subject

23 Z Re: Meeting

24 X Meeting

26 Y Re: Meeting
 Distrubuted 2.18
send receiv receiv Systems
X e e
1 m 4
1 m
sen
2 receiv
2 d Physic
Y e
receiv 3 al time
e
sen
Z d
receive

m3 m1 m2
A
receive receive
t t
1 2
receive t3
Fig 2.6: Ordering of events
2.2.3.2 Failure Model
In a DS, both processes and communication channels may fail – i.e., they may depart
from what is considered to be correct or desirable behavior. The following are the common
types of failures:
 Omission Failure

 Arbitrary Failure
 Timing Failure

a) Omission failure
Omission failures occur due to communication link failures. They are detected through
timeouts. They are classified as:

 Process omission failures

 Omission failures that occur due to process crash (i.e.) the execution of
the process could not continue.
 This is detected using timeouts.

 A fixed period of time is fixed for all the methods to complete its
execution. If the method takes time longer than the allowed time, a time
out has occurred.
 In an asynchronous system a timeout can indicate only that a process is
not responding.
 A process crash is called fail-stopif other processes can detect certainly
that the process has
crashed.
 Distrubuted 2.19
Systems
 Fail-stop behavior can be produced in a synchronous system if the
processes use timeouts to detect when other processes fail to respond and
messages are guaranteed to be delivered.

 Communication omission failures

 This occurs when the messages are dropped between sender and the
receiver.
 The message can be lost in sender buffer, receiver buffer or even in the
communication channel.
 The loss of messages between the sending process and the outgoing
message buffer is known as send omission failures.
 The loss of messages between the incoming message buffer and the
receiving process as receive-omission failures.

 The loss of messages in a channel is channel omission failure.

process p process q
send m receive

Communication
channel
Outgoing message buffer Incoming message buffer

Fig 2.7: Communication between processes in a channel

b) Arbitrary failures (Byzantine failure):
 This includes all possible errors that could cause failure.
 Communication channels often suffer from arbitrary failures
 The following table gives an overview of the different failures and its
categories:

Class Affects Comments

Fail stop Process Process halts and remains halted. Other
processes may detect that this process has
halted.
Crash Process Process halts and remains halted. Other
processes may not detect that this process has
halted.
 Distrubuted 2.20
Systems
Omission Channel A message inserted in an outgoing message
buffer never arrives at the other end‟s
incoming messagebuffer.
Send-omission Process A process completes a send operation but the
message is not put in its outgoing message
buffer.
Receiveomission Process A message is put in a process‟s incoming
messagebuffer, but that process does not
receive it.
Arbitrary Process Process/channel exhibits arbitrary behaviour:
or it may send/transmit arbitrary messages at
channel arbitrary times or commit omissions; a
process may stop or take an incorrect step.
c) Timing failures
 Timing failures refer to a situation where the environment in which a system
operates does not behave as expected regarding the timing assumptions, that is,
the timing constraints are not met.
 Timing failures are applicable in synchronous distributed system where time
limits are set on process execution time, message delivery time and clock drift
rate.
 The following are the common failures withnrespect to timings:

Class Affects Comments

Clock Process Process‟s local clock exceeds the bounds on its rate of drift
from real time

Performance Process Process exceeds the bounds on the interval between two steps

Performance Channel A message‟s transmission takes longer than the stated bound.

Masking failures
 Each component in a distributed system is generally constructed from a collection
of other components. It is always possible to construct reliable services from
components that exhibit failures.
 A knowledge of failure characteristics of a component can enable a new service to
be designed to mask the failure of the components on which it depends.
 Distrubuted 2.21
Systems
Reliability of one-to-one communication
The term reliable communication is defined in terms of validity and integrity.
 Validity: Any message in the outgoing message buffer is eventually delivered
to the incoming message buffer.
 Integrity: The message received is identical to one sent, and no messages are
delivered twice.
The threats to integrity come from two independent sources:
 Any protocol that retransmits messages but does not reject a message that
arrives twice.
 Malicious users that may inject spurious messages, replay old messages or
tamper with messages.
2.2.3.2 Security model

The security of a DS can be achieved by securing the processes and the channels
used in their interactions and by protecting the objects that they encapsulate
against unauthorized access.

Protection is described in terms of objects, although the concepts apply equally well to
resources of all types.

Protecting objects

Fig 2.8: Objects and Principals

 The server manages a collection of objects on behalf of some users.
 The users can run client programs that send invocations to the server to perform
operations on the objects.
 The server carries out the operation specified in each invocation and sends the
result to the client.
 Distrubuted 2.22
Systems
 This uses use “access rights” that define who is allowed to perform operation on a
object.

 The server should verify the identity of the principal (user) behind each operation
and checking that they have sufficient access rights to perform the requested
operation on the particular object, rejecting those who do not.
 A principal is an authority that manages the access rights. The principal may be a
user or a process.

Securing processes and their interactions

 The messages send by the processes are exposed to attack because the network and
the communication service that they use are open.
 To model security threats, we postulate an enemy that is capable of se nding any
process or reading/copying message between a pair of processes
 Threats form a potential enemy includes threats to processes, t hreats to
communication channels, and denial of service.

Fig: 2.9: Enemy

 Threats to processes:
 A process that is designed to handle incoming requests may receive a message
from any other process in the distributed system
 It cannot necessarily determine the identity of the sender.
 This lack of reliable knowledge of the source of a message is a threat to the
correct functioning of both servers and clients.
 Servers:
 Servers cannot necessarily determine the identity of the principal behind any
particular invocation.
 Without reliable knowledge of the sender‟s identity, a server cannot tell
whether to perform the operation or to reject it.
 Distrubuted 2.23
Systems
 Clients:
 When a client receives the result of an invocation from a server, it cannot
necessarily tell whether the source of the result message is from the intended
server or from an enemy, perhaps „spoofing‟ the mail server.
 Thus the client could receive a result that was unrelated to the original
invocation, such as a false mail item (one that is not in the user‟s mailbox).
 Threats to communication channels:
 An enemy can copy, alter or inject messages as they travel across the network
and its intervening gateways.
 Such attacks present a threat to the privacy and integrity of information as it
travels over the network and to the integrity of the system.
 Defeating security threats
 Encryption and authentication are use to build secure channels.
 Each of the processes knows the identity of the principal on whose behalf the
other process is executing and can check their access rights before performing
an operation.
Other possible threats from an enemy
 Denial of service:
Denial of service (DoS) attack is an incident in which a user or organization is
deprived of the services of a resource they would normally expect to have. In a
distributed denial-of-service, large numbers of compromised systems (sometimes
called a botnet) attack a single target.
 Mobile code:
Mobile code is software transferred between systems, e.g. transferred across a
network, and executed on a local system without explicit installation by the
recipient. Mobile code raises new and interesting security problems for anyprocess
that receives and executes program code from elsewhere.
The uses of security models
The use of security techniques incurs processing and management costs. The
distributed systems faces threats from various points. The threat analysis demands the
construction of security models.
2.3 INTERPROCESS COMMUNICATION
Interprocess communication (IPC) is a set of programming interfaces that allow
a programmer to coordinate activities among different program processes that
can run concurrently in an operating system.
 Distrubuted 2.24
Systems
- IPC refers to exchange of data between two or more separate, independent
processes/threads.
- The operating systems provide facilities/resources for inter-process
communications (IPC), such as message queues, semaphores, and shared memory.
- Distributed computing systems make use of these facilities/resources to provide
application programming interface (API) which allows IPC to be programmed at a
higher level of abstraction. (e.g., send and receive)
- Distributed computing requires information to be exchanged among independent
processes.
2.3.1 The API’s for the Internet
Protocols Characteristics of IPC:

 Synchronous and asynchronous communication

 In the synchronous form of communication, the sending and receiving
processes synchronize at every message.
 In this case, both send and receive are blocking operations.
 In the asynchronous form of communication, the use of the send operation is
non blocking in that the sending process is allowed to proceed as soon as the
message has been copied to a local buffer, and the transmission of the message
proceeds in parallel with the sending process.
 The receive operation can have blocking and non-blocking variants.

 Message destinations
 The messages in the Internet protocols, messages are sent to (Internet address,
local port) pairs.
 A local port is a message destination with in a computer, specified as an
integer.
 A port has exactly one receiver but can have many senders.
 Processes may use multiple ports to receive messages.
 Any process that knows the number of a port can send a message to it.
 Client programs refer to services by name and use a name server or binder to
translate their names into server locations at runtime.
 This allows services to be relocated but not to migrate – that is, to be moved
while the system is
running.
 Distrubuted 2.25
Systems
 Reliability
 A point-to-point message service can be described as reliable if messages are
guaranteed to be delivered.
 A point-to-point message service can be described as unreliable if messages are
not guaranteed to be delivered in the face of even a single packet dropped or
lost.
 Ordering
 Some applications require that messages be delivered in the order in which
they were transmitted by the sender.
 The delivery of messages outof sender order is regarded as a failure by such
applications.
Sockets
A socket is one endpoint of a two-way communication link between two programs
running on the network.

Interprocess communication consists of transmitting a message between a socket in

one process and a socket in another process. The java API for interprocess communication in
the internet provides both datagram and stream communication. Both forms of
communication, UDP and TCP, use the socket abstraction, which provides and endpoint for
communication between processes. Interprocess communication consists of transmitting a
message between a socket in one process and a socket in another process. The two
communication patterns that are most commonly used in distributed programs:
 Client-Server communication
 The request and reply messages provide the basis for remote method
invocation (RMI) or remote procedure call (RPC).
 Group communication
 The same message is sent to several processes.

Fig 2.6: Sockets and ports

 Distrubuted 2.26
Systems
For a process to receive messages, its socket must be bound to a local port and one of
the Internet addresses of the computer on which it runs. Messages sent to a particular Internet
address and port number can be received only by a process whose socket is associated with
that Internet address and port number.

Java API for Internet addresses

 Java provides a class, InetAddress, that represents Internetaddresses.

 Users of this class refer to computers by Domain Name System (DNS) hostnames.
InetAddressaComputer = InetAddress.getByName("abc.ac.in");

UDP Datagram Communication

 The application program interface to UDP provides a message passing abstraction.

 Message passing is the simplest form of interprocess communication.

 API enables a sending process to transmit a single message to a receiving process.

 The independent packets containing theses messages are called datagrams.

 In the Java and UNIX APIs, the sender specifies the destination using a socket.

 A datagram is transmitted between processes when one process sends it and

another receives it.

 The following are the steps in socket communication:

o Creating a socket

o Binding a socket to a port and local Internet address

 A client binds to any free local port

 A server binds to a server port

 Receive() returns Internet address and port of sender, along with the message.

Issues in datagram communication:

 Message size:
 The receiving process needs to specify an fixed size array to receive a message.
 If the message is too big for the array, it is truncated.

 Any application requiring messages larger than the maximum must fragment
them.
 Distrubuted 2.27
Systems
 Blocking:

 Sockets normally provide non-blocking sends and blocking receives for

datagram communication.

 The send() returns after delivering the message to theUDP and IP protocols.
These protocols are now responsible for transmitting the message to its
destination

 The message is placed in a queue for the socket that is bound to the destination
port.

 Messages are discarded atthe destination if no process has a socket bound to

the destination port.

 The receive()blocks until a datagram is received, unless a timeout has been set
on the socket.

 Timeouts:

 The receive() cannot wait indefinitely. This situation occurs when the sending
process may have crashed or the expected message may have been lost.

 To avoid this timeouts can be set on sockets.

 The timeouts must be much larger than the time required to transmit a
message.

 Receive from any:

 The receive() does not specify an origin for messages. This can get datagrams
addressed to its socket from any origin.

 The receive() returns the Internet address and local port of the sender, allowing
the recipient to check where the message came from.

 It is possible to connect a datagram socket to a particular remote port and

Internet address, in which case the socket is only able to send messages to and
receive messages from that address.

Failure model for UDP datagrams

UDP datagrams suffer from following failures:

 Omission failure: Messages may be dropped occasionally

 Ordering: Messages can be delivered out of order.

 Distrubuted 2.28
Systems
Java API for UDP datagrams
The Java API provides datagram communication by two classes:
1. DatagramPacket
- It provides a constructor to make an array of bytes comprising:
• Message content
• Length of message
• Internet address
• Local port number
- It provides another similar constructor for receiving a message.
2. DatagramSocket
- This class supports sockets for sending and receiving UDP datagram.
- It provides a constructor with port number as argument.
- No-argument constructor is used to choose a free local port.
- DatagramSocket methods are:
1. send and receive:These methods are used for transmitting
datagrams between a pair of sockets.

Syntax:
send(packet);
receive();
2. setSoTimeout: This method allows a timeout to be set. With a
timeout set, the receive method will block for the time specified and
then throw an InterruptedIOException.

Syntax:
setSotimeout(time in seconds);
3. connect: This method is used for connecting to a particular remote
port and Internet address, in which case the socket is only able to
send messages to and receive messages from that address.

Syntax:
int connect(in sockfd, const structsockaddr *addr, socklen_taddrlen);
 Distrubuted 2.29
Systems
Arrays of bytes Message Internet Port number
containing message length address

Fig 2.7: Datagram

Packet UDP Client server Communication
Client Program

import java.net.*;
import java.io.*;
public class UDPClient
{
public static void main(String args[])
{
DatagramSocketaSocket = null;
try {
aSocket = new DatagramSocket(); // Create an object for the datagram class
byte [] m = args[0].getBytes(); //Buffer to get data
InetAddressaHost = InetAddress.getByName(args[1]); // Get Inetaddress
intserverPort = 6789;
DatagramPacket request =
new DatagramPacket(m, m.length(), aHost, serverPort);
aSocket.send(request); //Sent the packet
byte[] buffer = new byte[1000];
DatagramPacket reply = new DatagramPacket(buffer, buffer.length);
aSocket.receive(reply);
System.out.println("Reply: " + new String(reply.getData()));
}
catch (SocketException e){System.out.println("Socket: " + e.getMessage());
}
catch (IOException e){System.out.println("IO: " + e.getMessage());
}
 Distrubuted 2.30
Systems
finally
{
if(aSocket != null) aSocket.close();}
}
}

Server program
import java.net.*;
import java.io.*;
public class UDPServer
{
public static void main(String args[])
{
DatagramSocketaSocket = null;
try
{
aSocket = new DatagramSocket(6789);
byte[] buffer = new byte[1000];
while(true)
{
DatagramPacket request = new DatagramPacket(buffer, buffer.length);
aSocket.receive(request);
DatagramPacket reply = new DatagramPacket(request.getData(),
request.getLength(), request.getAddress(), request.getPort());
aSocket.send(reply);
}
}
catch (SocketException e)
{
System.out.println("Socket: " + e.getMessage());
 Distrubuted 2.31
Systems
}
catch (IOException e)
{
System.out.println("IO: " + e.getMessage());
}
finally
{
if (aSocket != null) aSocket.close();}
}
}

TCP stream Communication

The API to the TCP protocol provides the abstraction of a stream of bytes to be written
to or read from. The following features are hidden stream abstraction:
 Message sizes
 Lost messages
 Flow control
 Message destinations
Working of stream communication
 The pair of sockets in the client and server are connected by a pair of streams, one
in each direction. Thus each socket has an input stream and an output stream.
 One of the pair of processes can send information to the other by writing to its
output stream, and the other process obtains the information by reading from its
input stream.
 When an application closes a socket, this indicates that it will not write any more
data to its output stream. Any data in the output buffer is flushed to the other end
of the stream and put in the queue at the destination socket, with an indication that
the socket is closed.
 The process at the destination can read the data in the queue, but any further reads
after the queue is empty will result in an indication of end of stream.
 When a process exits or fails, all of its sockets are eventually closed and any
process attempting to communicate with it will discover that its connection has
been
broken.
 Distrubuted 2.32
Systems
This form of stream communication has the following issues:
 Matching of data items: Two communicating processes need to agree as to
the contents of the data transmitted over a stream.
 Blocking: The process that writes data to a stream may be blocked by the TCP
flow-control mechanism if the socket at the other end is queuing as much data
as the protocol allows.
 Threads: When a server accepts a connection, it generally creates a new thread
in which to communicate with the new client. In an environment in which
threads are not provided, an alternative is to test whether input is available
from a stream before attempting to read it;
Java API for TCP streams
The Java interface to TCP streams is provided in the classes:
 ServerSocket: It is used by a server to create a socket at server port to listen
for connect requests from clients.
Syntax:
publicServerSocket(int port, InetAddress address) ;
 Socket: It is used by a pair of processes with a connection. The client uses a
constructor to create a socket and connect it to the remote host and port of a
server.It provides methods for accessing input and output streams associated
with a socket.
Syntax:
public Socket(InetAddress host, int port, InetAddresslocalAddress,
intlocalPort);

TCP Client Server Communication

TCP Client
import java.net.*;
import java.io.*;
public class TCPClient
{
public static void main (String args[])
{
Socket s = null;
 Distrubuted 2.33
Systems
try
{
intserverPort = 7896;
s = new Socket(args[1], serverPort);
DataInputStream in = new DataInputStream( s.getInputStream());
DataOutputStream out = new DataOutputStream( s.getOutputStream());
out.writeUTF(args[0]); // UTF is a string encoding;
String data = in.readUTF();
System.out.println("Received: "+ data)
;
}
catch (UnknownHostException e)
{
System.out.println("Sock:"+e.getMessage());
}
catch (EOFException e)
{ System.out.println("EOF:"+e.getMessage());
}
catch (IOException e)
{ System.out.println("IO:"+e.getMessage());
}
finally
{ if(s!
=null) try
{
s.close();
}
catch (IOException e)
{
 Distrubuted 2.34
Systems
/*close failed*/
}
}
}
}

TCP server
import java.net.*;
import java.io.*;
public class TCPServer
{
public static void main (String args[])
{
try{
intserverPort = 7896;
ServerSocketlistenSocket = new ServerSocket(serverPort);
while(true) {
Socket clientSocket = listenSocket.accept();
Connection c = new Connection(clientSocket);
}
} catch(IOException e) {System.out.println("Listen :"+e.getMessage());}
}
}
class Connection extends Thread
{
DataInputStream in;
DataOutputStream out;
Socket clientSocket;
 Distrubuted 2.35
Systems
public Connection (Socket aClientSocket)
{
try {
clientSocket = aClientSocket;
in = new DataInputStream( clientSocket.getInputStream());
out =new DataOutputStream( clientSocket.getOutputStream());
this.start();
} catch(IOException e) {System.out.println("Connection:"+e.getMessage());}
}
public void run(){
try { // an echo server
String data = in.readUTF();
out.writeUTF(data);
} catch(EOFException e) {System.out.println("EOF:"+e.getMessage());
} catch(IOException e) {System.out.println("IO:"+e.getMessage());
} finally { try {clientSocket.close();}catch (IOException e){/*close failed*/}}
}
}

2.4 EXTERNAL DATA REPRESENTATION AND MARCHALLING

The information stored in running programs is represented as data structures, whereas
the information in messages is in the form of sequences of bytes. Irrespective of the form of
communication used, the data structure must be converted to a sequence of bytes before
transmission and must be rebuilt on arrival.

External Data Representation is an agreed standard for the representation of data

structures and primitive values.

But there are some problems in data representation. They are:

 Two conversions are necessary (i.e.) conversion of bytes to data structures and
vice versa in case of using an agreed format at both ends.
 Using sender‟s or receiver‟s format and convert at the other end.
 Distrubuted 2.36
Systems
Marshalling is the process of taking a collection of data items and assembling
them into a form suitable for transmission in a message. Unmarshalling is the
process of disassembling a collection of data on arrival to produce an equivalent
collection of data items at the destination.

The following are the three common approaches to external data representation and
marshalling:
 CORBA: The data representation is concerned with an external representation for
the structured and primitive types that can be passed as the arguments and results
of remote method invocations in CORBA. It can be used by a variety of
programming languages. Marshalling and unmarshalling is done by the
middleware in binary form.
 Java’s object serialization: This is concerned with the flattening and external data
representation of any single object or tree of objects that may need to be
transmitted in a message or stored on a disk. It is for use only by Java. Marshalling
and unmarshalling is done by the middleware in binary form.
 XML: This defines a textual fomat for representing structured data. It was
originally intended for documents containing textual self-describing structured
data. Marshalling and unmarshalling is in texual form.

2.4.1 CORBA’a Common Data Representation (CDR)

CORBA CDR is the external data representation defined with CORBA 2.0.It consists
15 primitive types:
 Short (16 bit)

 Long (32 bit)

 Unsigned short
 Unsigned long
 Float(32 bit)
 Double(64 bit)
 Char
 Boolean(TRUE,FALSE)
 Octet(8 bit)
 Any(can represent any basic or constructed type)
 Distrubuted 2.37
Systems
There are two types of CDR in CORBA:
 Primitive types: CDR defines a representation for both big-endian and little-
endian orderings. Characters are represented by a code set agreed between
client and server.
 Constructed types: The primitive values that comprise each constructed type
are added to a sequence of bytes in a particular order. The following are the
constructed types:
 sequence length (unsigned long): followed by elements in order
 string length (unsigned long): followed by characters in order (can also
have wide characters)
 array: array elements in order (no length specified because it is fixed)
 struct: in the order of declaration of the components
 enumerated unsigned long: the values are specified by the order declared
 union: type tag followed by the selected member
Example:
struct with value {„Sona‟, „Chennai‟, 1986}

0-3 4 Length of the string

4-7 “Sona” String
8-11 “@ ”
12-15 7 Length of the string
16-19 “Chen” Chennai
20-23 “on ”
24-27 1986 Unsigned long

Marshalling in CORBA
The type of a data item not given in case of CORBA. CORBA assumes that both the
sender and recipient have common knowledge of the order and types of data items. The types
of data structures and types of basic data items are described in CORBA IDL. This provides a
notation for describing the types of arguments and results of RMI methods
 Distrubuted 2.38
Systems
Example:
Struct student
{
string name;
stringrollnumber;
}

2.4.2 Java Object Serilaization

In Java RMI, both object and primitive data values may be passed as arguments and
results of method invocation. An object is an instance of a Java class. The Java equivalent for
the above mentioned CORBA construct is :
Public class student implements Serializable
{ Private String name;
Private String
rollnumber;
Public student(String aName ,String arollnumber)
{
name = aName;
rollnumber= arollnumber;
}
}

Student 8 bit version number h0 Class name, version

2 Java.lang.string.name Java.lang.string.rollnumber Number, type and

name of instance
variables

5Sona 8CS50 h1 Values of variables

To serialize an object: ‹
(1) its class info is written out: name, version number ‹
(2) types and names of instance variables • If an instance variable belong to a new
class, then new class info must be written out, recursively. Each class is given a
handle
‹
 Distrubuted 2.39
Systems
(3) values of instance variables ‹
Example: student s = new student(“5Sona”, “8CS50”);
‹ Thefollowinga re the steps to serialize objects in Java:
 create an instance of ObjectOutputStream ‹
 Invoke writeObject method passing Person object as argument ‹
To deserialize
 create an instance of ObjectInputStream ‹
 Invoke readObject method to reconstruct the original
object ObjectOutputStream out = new
ObjectOutputStream(… );
out.writeObject(originalStudent);
ObjectInputStream in = new ObjectInputStream(…);
Student theStudent = in.readObject();
Reflection:
Reflection is inquiring about class properties, e.g., names, types of methods and
variables, of objects. This allows to perform serialization and deserialization in a generic
manner, unlike in CORBA, which needs IDL specifications. For serialization, use reflection to
find out
(1) class name of the object to be serialized
(2) the names, types
(3) values of its instance variables
„ For deserialization,
(1) class name in the serialized form is used to create a class
(2) it is then used to create a constructor with arguments types corresponding to those
specified in the serialized form.
(3) the new constructor is used to create a new object with instance variables whose
values are read from the serialized form

2.4.3 XML (Extensible Markup Language)

 XML is a markup language that was defined by the World Wide Web Consortium
(W3C) for general use on the Web.
 Markup language refers to a textual encoding that represents both a text
and details as to its structure or its appearance.
 Distrubuted 2.40
Systems
 In XML, the data items are tagged with „markup‟ strings.
 The tags are used to describe the logical structure of the data and to associate
attribute-value pairs with logical structures.
 XML is used to enable clients to communicate with web services and for defining
the interfaces and other properties of web services.
 XML is extensible in the sense that users can define their own tags, in contrast to
HTML, which uses a fixed set of tags.
 XML could be used by multiple applications for different purposes.
Example:
<student id="123456789">
<name>Sona</name>
<rollnumber>CS50</rollnumber>
<year>1984</year>

XML elements and attributes

 Elements: An element in XML consists of a portion of character data surrounded
by matching start and end tags.
 Attributes: A start tag may optionally include pairs of associated attribute names
and values such as id="123456789".
 Names: The names of tags and attributes in XML generally start with a letter, but
can also start with an underline or a colon. The names continue with letters, digits,
hyphens, underscores, colons or full stops. Letters are case-sensitive. Names that
start with xml are reserved.
 Binary data: All of the information in XML elements must be expressed as
character data. The encrypted elements or secure hashes are represented as base64
notation.
 CDATA: XML parsers normally parse the contents of elements because they may
contain further nested structures. If text needs to contain an angle bracket or a
quote, it may be represented in a special way.
 Namespace: A namespace applies within the context of the enclosing pair of start
and end tags unless overridden by an enclosed namespace declaration.
2.4.3 Remote Object Reference
 Remote object references are used when a client invokes an object that is located
on a remote
server.
 Distrubuted 2.41
Systems
 A remote object reference is passed in the invocation message to specify which
object is to be invoked.
 Remote object references must be unique over space and time.

Fig 2.8: Representation of remote object references

The following are mandatory for remote object reference
 internet address/port number: process which created object

 time: creation time

 object number: local counter, incremented each time an object is created in
the creating process
 interface: how to access the remote object (if object reference is passed from
one client to another)

2.5 MULTICAST COMMUNICATION

Multicast (one-to-many or many-to-many distribution) is group

communication where information is addressed to a group of destination
computers simultaneously.

Multicast operation is an operation that sends a single message from one process to
each of the members of a group of processes. The simplest way of multicasting, provides no
guarantees about message delivery or ordering.
The following are the characteristics of multicasting:
 Fault tolerance based on replicated services:
 A replicated service consists of a group of servers.
 Client requests are multicast to all the members of the group, each of which
performs an identical operation.
 Finding the discovery servers in spontaneous networking

 Multicast messages can be used by servers and clients to locate available

discovery services in order to register their interfaces or to look up the
interfaces of other services in the distributed system.
 Distrubuted 2.42
Systems
 Better performance through replicated data
 Data are replicated to increase the performance of a service.
 Propagation of event notifications

 Multicast to a group may be used to notify processes when something happens.

Join to receive
multicast message

Receiver

Multicast IP
Address

Sender Receiver

Receiver
Fig 2.9: Multicasting
2.5.1 IP Multicast
 IP multicast is built on top of the Internet protocol.
 IP multicast allows the sender to transmit a single IP packet to a multicast group.
 A multicast group is specified by class D IP address with1110 as it starting byte.
 The membership of a multicast group is dynamic.
 A computer is said to be in a multicast group if one or more processes have
sockets that belong to the multicast group.
The following details are important when using IPV4 IP multicasting:

 Multicast IP routers
 IP packets can be multicast both on local network and on the Internet.
 Local multicast uses local network such as Ethernet.
 To limit the distance of propagation of a multicast datagram, the sender can
specify the number of routers it is allowed to pass- called the time to live
(TTL).
 Distrubuted 2.43
Systems
 Multicast address allocation

 Multicast addressing may be permanent or temporary.

 Permanent groups exist even when there are no members.

 Multicast addressing by temporary groups must be created before use and will
stop to exit when all members have left the particular multicast group.

 To start or to join a particular multicast group, a session directory (sd)

program is used.

 This is a tool with an interactive interface that allows users to browse

advertised multicast sessions and to advertise their own session, specifying the
time and duration that it wishes to be in the group.

Failure model for IP multicast datagram

They suffer from omission failures. The messages are not guaranteed to be delivered
to the group member in case of omission failures.

2.5.2 Java API to IP multicast

 The Java API provides a datagram interface to IP multicast through the class
MulticastSocket.

 MulticastSocket is a subset of DatagramSocket.

 It has the capability of being able to join multicast groups.

 The multicast datagram socket class is useful for sending and receiving IP
multicast packets.

 A MulticastSocket is a (UDP) DatagramSocket, with additional capabilities for

joining "groups" of other multicast hosts on the internet.

 The class MulticastSocket provides two constructors:

 MulticastSocket(): Create a multicast socket.

 MulticastSocket(int): Create a multicast socket and bind it to a specific port.

MulticastSender.java
import java.io.*;
import java.net.*;
public class MulticastSender {
 Distrubuted 2.44
Systems
public static void main(String[] args) {
DatagramSocket socket = null;
DatagramPacketoutPacket = null;
byte[] outBuf;
final int PORT = 8888;
try {
socket = new DatagramSocket();
long counter = 0;
String msg;
while (true) {
msg = " multicast! " + counter;
counter++;
outBuf = msg.getBytes();
//Send to multicast IP address and port
InetAddress address = InetAddress.getByName("224.2.2.3");
outPacket = new DatagramPacket(outBuf, outBuf.length, address, PORT);
socket.send(outPacket);
System.out.println("Server sends : " + msg);
try {
Thread.sleep(500);
} catch (InterruptedExceptionie) {
}
}
} catch (IOExceptionioe) {
System.out.println(ioe);
}
}
}
 Distrubuted 2.45
Systems
Multicastreceiver.java
import java.io.*;
import java.net.*;
public class MulticastReceiver {
public static void main(String[] args) {
MulticastSocket socket = null;
DatagramPacketinPacket = null;
byte[] inBuf = new byte[256];
try {
//Prepare to join multicast group
socket = new MulticastSocket(8888);
InetAddress address = InetAddress.getByName("224.2.2.3");
socket.joinGroup(address);
while (true) {
inPacket = new DatagramPacket(inBuf, inBuf.length);
socket.receive(inPacket);
String msg = new String(inBuf, 0, inPacket.getLength());
System.out.println("From " + inPacket.getAddress() + " Msg : " + msg);
}
} catch (IOExceptionioe) {
System.out.println(ioe);
}
}
}

The following methods are available in MulticastSocket class:

 getInterface():Retrieve the address of the network interface used for multicast
packets.
 getTTL():Get the default time-to-live for multicast packets sent out on the socket.
 Distrubuted 2.46
Systems
 joinGroup(InetAddress): Joins a multicast group.
 leaveGroup(InetAddress):Leave a multicast group.
 send(DatagramPacket, byte):Sends a datagram packet to the destination, with a
TTL (time- to-live) other than the default for the socket.
 setInterface(InetAddress): Set the outgoing network interface for multicast
packets on this socket, to other than the system default.
 setTTL(byte): Set the default time-to-live for multicast packets sent out on this
socket.

2.5.2 Reliability and ordering in multicasting

There are many factors that puts the reliability of multicasting under question:
 A datagram sent from one multicast router to another may be lost.
 There are chances for a single datagram to arrive at multiple recipients, any one of
those recipients may drop the message because its buffer is full.
 If a multicast router fails, the group members beyond that router will not receive
the multicast message.

Ordering issue:
IP packets do not necessarily arrive in the order in which they were sent. Messages
sent by two different processes will not necessarily arrive in the same order at all the members
of the group. The problem of reliability and ordering is more common in:
 Fault tolerance based on replicated services
 Discovering services in spontaneous networking
 Better performance through replicated data
 Propagation of event notifications

2.6 NETWORK VIRTULIZATION: OVERLAY NETWORKS

Network virtualization is the construction of many different virtual networks over an
existing network such as the Internet.

Network virtualization refers to the management and monitoring of an entire

computer network as a single administrative entity from a single software-
based administrator’s console.
 Distrubuted 2.47
Systems
NV is implemented by installing software and services to manage the sharing
of storage, computing cycles and applications. All servers and services in the network are
treated as a single pool of resources that can be accessed without regard for
its physical components.

2.6.1 Overlay Networks

An overlay network is a virtual network of nodes and logical links that is built on
top of an existing network with the purpose to implement a network service that is
not available in the existing network.

The overlay network may sometimes change the properties of the underlying network.
The overlay network provides:
 a service that is tailored towards the needs of a class of application or a particular
higher-level service
 more efficient operation in a given networked environment
 additional feature – for example, multicast or secure communication.
Advantages:
 They enable new network services to be defined without requiring changes to the
underlying network.
 They encourage experimentation with network services and the customization of
services to particular classes of application.
 Multiple overlays can be defined and can coexist, with the end result being a more
open and extensible network architecture.

Disadvantages:
 They introduce an extra level of indirection.
 They add to the complexity of network services.

Types of Overlay Networks

The overlay networks are classified based on the following criteria:

1. Classification based on application:

 Distributed hash tables:
A method for storing hash tables in geographically distributed locations in
order to provide a failsafe lookupmechanism for distributed computing.
The
 Distrubuted 2.48
Systems
overlay network offers a service that manages a mapping from keys to values
across a potentially large number of nodes in a completely decentralized
manner (similar to a standard hash table but in a networked environment).

 Peer-to-peer file sharing:

Overlay structures that focus on constructing tailored addressing and routing
mechanisms to support the cooperative discovery and use of files.

 Content distribution networks:

Overlays that subsume a range of replication, caching and placement strategies
to provide improved performance in terms of content delivery to web users;
used for web acceleration and to offer the required real-time performance for
video streaming.

2. Classification based on network style:

 Wireless ad hoc networks: Network overlays that provide customized routing

protocols for wireless ad hoc networks.

 Disruption-tolerant networks: Overlays designed to operate in hostile

environments that suffer significant node or link failure and potentially high
delays

3. Classification based on additional features:

 Multicast: Overlay networks provide access to multicast services where

multicast routers are not available.
 Security: Overlay networks that offer enhanced security over the underling IP
network, including virtual private networks.
 Resilience: Overlay networks that seek an order of magnitude improvement in
robustness and availability of Internet paths

2.6.2 Skype –An overlay Network

Skype employs a partially decentralized architecture – a mix of the peer-to-peer and
client-server architectures. The client-server system is used for authentication while the peer-
to-peer system is used for IP telephony, relaying, indexing peers and file transfers. On top of
this is the Skype overlay network which users interact with directly, an overlay network is a
virtual network that is formed above and independent of the underlying Internet protocol
network. Skype uses overlay networks to achieve the following:
1. It enables them to design and use their own proprietary protocols and
application over the Internet. This enables Skype to encrypt all packet
transmissions.
 Distrubuted Systems 2.4
9
2. Redundant links between Skype peers guarantees robustness. This assures of e
to-end connection even when network resources are severely limited. nd-
3. The peer-to-peer overlay model provides an elegant solution to the
scalability challenge, given the huge number of interconnected nodes.
4. Skype uses its overlay network to circumvent problems arising rom
f Network Address Translation traversal and bypass firewalls.
5. Routing and relaying are very flexible; protocols allow peers to autonomou sly
set up links with the best nodes available.
6. The overlay allows Skype to store a lot of offline data in its peers lea ding
to reduced storage infrastructure expenses.
7. Employing peer relays improves the quality of VoIP calls.

Fig 2.10: Skype Overlay Network

Skype architecture
 Skype is based on a peer-to-peer infrastructure consisting of ordinary users‟
machines called hosts and super nodes.
 Super nodes are ordinary Skype hosts that happen to have sufficient capabilities to
carry out their enhanced role.
 Super nodes are selected on demand based a range of criteria including bandwidth
available, reach ability and availability.
 Distrubuted 2.50
Systems
User connection

 Skype users are authenticated through a login server.

 They then make contact with a selected super node. To achieve this, each client
maintains a cache of super node identities.

 At first login this cache is filled with the addresses of around seven super nodes,
and over time the client builds and maintains a much larger set (perhaps several
hundred).

Search for users

 Super nodes search the global index of users, which is distributed across the super
nodes.

 The search is orchestrated by the client‟s chosen super node and involves an
expanding search of other super nodes until the specified user is found.

 A user search typically takes between three and four seconds to complete for hosts
that have a global IP address.

Voice connection

 Skype establishes a voice connection between the two parties using TCP for
signaling all requests and terminations and either UDP or TCP for the streaming
audio.

2.7 MESSAGE PASSING INTERFACE (MPI)

Message Passing Interface (MPI) is a standardized and portable message-passing

system to function on a wide variety of parallel computers.

 MPI primarily addresses the message-passing parallel programming model: data is

moved from the address space of one process to that of another process through
cooperative operations on each process.
 The primary goal of the Message Passing Interface is to provide a widely used
standard for writing message passing programs.
 The interface attempts to be: practical, portable, efficient and flexible.
 The basic principles of exchanging messages between two processes using send
and receive operations is covered in
MPI.
 Distrubuted 2.51
Systems
 There are two types of message passing:
 Synchronous message passing: This is implemented by blocking send and
receive calls.
 Asynchronous message passing: This is implemented by a non-blocking
form of send.
Originally, MPI was designed for distributed memory architectures.As architecture
trends changed, shared memory SMPs were combined over networks creating hybrid
distributed memory / shared memory systems. MPI runs on virtually any hardware platform
like distributed memory, shared memory and in hybrid memory.

Send Operations Blocking Mode Non Blocking Mode

Generic MPI_Send: MPI_Isend:
the sender blocks until it is safe to the call returns immediately and
return (i.e.) until the message is in the programmer is given a
transit or delivered and the communication request handle,
sender‟s application buffer can which can then be used to check
therefore be reused. the progress of the call via
MPI_WaitorMPI_Test.
Synchronous MPI_Ssend: MPI_Issend:
the sender and receiver as with MPI_Isend,but with
synchronize and the call only MPI_WaitandMPI_Testindicating
returns when the message has been whether the message has been
delivered at the receiving end. delivered at the receive end.
Buffered MPI_Bsend: MPI_Ibsend:
the sender explicitly allocates an as with MPI_I send but with
MPI buffer library and the call MPI_WaitandMPI_Testindicating
returns when the data is whether the message has been
successfully copied into this copied into the sender‟s MPI
buffer. buffer and hence is in transit.
Ready MPI_Rsend: MPI_Irsend:
the call returns when the sender‟s the effect is as with MPI_Isend,
application buffer can be reused but as with MPI_Rsend, the
but the programmer is also programmer is indicating to the
indicating to the library that the underlying implementation that
receiver is ready to receive the the receiver is guaranteed to be
message, resulting in potential ready to receive.
optimization of the underlying
implementation.
2.52 Communication in Distrubuted Sytem Distrubuted 2.52
Systems
2.8 REMOTE INVOCATION
In a distributed object system, an object's data should be accessible only via its
methods. The objects can be accessed via object references. The following terminologies are
important in remote invocations:
 An interface provides a definition of the signatures of a set of methods without
specifying their implementation.
 An action is initiated by an object invoking a method in another object.

 The state of an object consists of the values of its instance variables.

Remote invocations are method invocations for objects in different processes.

 Remote objects are objects that can receive remote

invocation. The core part of the distributed object model is:
- remote object reference;
- remote interface: specifies which methods can be invoked remotely;
2.8.1 Request reply protocols
The overview of request reply protocol is:
 The client sends a query to the server, and reads from the socket, usually in a
blocking way, for the server response.
 The server processes the command and sends the response back to the client.
The Remote Procedure Call (RPC) is a common model of request reply protocol.
There are two types of communication:
 Synchronous: The client process blocks until the reply arrives from the server.
It is termed as reliable form of communication since the reply from the server
is actually an acknowledgement to the client.
 Asynchronous: The clients can retrieve the replies from the server
later. UDP is preferred over TCP for the following reasons:
 Acknowledgements are redundant, since requests are followed by replies.
 Establishing a connection involves two extra pairs of messages in addition to
the pair required for a request and a reply.
 Flow control is redundant for the majority of invocations, which pass only
small arguments and results.
 Distrubuted 2.53
Systems
Request Reply protocol:
The following are the primitives in request reply protocols:
 public byte[] doOperation (RemoteObjectRef o, intmethodId, byte[]
arguments):
This method sends a request message to the remote object and returns the
reply. The arguments specify the remote object, the method to be invoked
and the arguments of that method.
 public byte[] getRequest ():
This method acquires a client request via the server port.
 public void sendReply (byte[] reply, InetAddressclientHost,
intclientPort):
This method sends the reply message reply to the client at its Internet
address and port.
Message identifiers
Each message have a unique message identifier for referencing it. A message identifier
consists of two parts:
 request Id: Increasing sequence of integers assigned by the sending
process. This makes the message unique to the sender.
 an identifier: for the sender process. This makes the message unique in the
distributed system.
Failure model of the request-reply protocol
The following are the common failures in request reply protocol:
 Omission failures.
 Messages are not guaranteed to be delivered in sender order.
 Failure of processes.
Discarding duplicate request messages
The server may receive the same request message more than once. To avoid this, the
protocol is designed to recognize successive messages from the same client with the same
request identifier and to filter out duplicates.
Lost reply messages
If the server has already sent the reply to which it receives a duplicate request it will
need to execute the operation again to obtain the result. This could be avoided if the server has
stored the result of the original execution.
 Distrubuted 2.54
Systems
History

History is used to refer to a structure that contains a record of reply messages that
have been transmitted by the server.

Each entry in the history has

 a request identifier
 a message
 client identifier
The main goal of history is to allow the server to retransmit replymessages when client
processes request them. The main issue faced by the history is the size. As more
communication takes place, more messages need to be saved in the history which incurs more
cost.
Styles of exchange protocols
There are three exchange protocols:
 request (R) protocol: a single request message is sent by the client to the server
 request-reply (RR) protocol: used in client-server exchanges because it is based
on the request-reply protocol. Special acknowledgement messages are not
required.
 request-reply-acknowledge reply (RRA) protocol: It is based on request-reply-
acknowledge reply. The Acknowledge reply message contains the requested from
the reply message being acknowledged. This will enable the server to discard
entries from its history. The arrival of a requested in an acknowledgement message
will be interpreted as acknowledging the receipt of all reply messages with lower
request Ids.
TCP streams to implement the request-reply protocol
It is difficult to determine the buffer size for the UDP datagrams. Also the fixed size of
the datagrams imposes a limit on the message size which is not favorable. So TCP streams
are preferred since they allow messages of any size to be transmitted. If the TCP protocol is
used, it ensures that request and reply messages are delivered reliably thereby avoiding
acknowledgments. The overhead due to TCP acknowledgement messages is reduced when a
reply message follows soon after a request message.
HTTP: An Example Request Reply protocol
Hyper Text Transfer Protocol (HTTP) is a stateless request-response based
communication protocol. It's used to send and receive data on the Web i.e., over the Internet.
This protocol uses reliable TCP connections either for the transfer of data to and from clients
(Web Browsers).
 Distrubuted 2.55
Systems

Fig 2.11: HTTP Request-Response protocol

HTTP specifies the messages involved in a request-reply exchange, the methods,
arguments and results, and the rules for representing (marshalling) them in the messages. This
protocol allows for content negotiation and password-style authentication
 Content negotiation: Clients‟ requests can include information as to what
data representations they can accept.
 Authentication: Credentials and challenges are used to support password-
style authentication.
The following are the connection establishment and closing steps:
 The client requests and the server accepts a connection at the default server
port or at a port specified in the URL.

 The client sends a request message to the server.

 The server sends a reply message to the client.
 The connection is closed after the communication is over.

Connections in HTTP
There are two types of connections in HTTP: Persistent and non persistent

Differences between persistent and non persistent connections

Persistent Connection Non persistent Connection

On the same TCP connection the server, The server parses request, responds, and
parses request, responds and waits for new closes TCP connection.
requests.

It takes fewer RTTs to fetch the objects. It takes 2 RTTs to fetch each object.

It has less slow start. Each object transfer suffers from slow start
 Distrubuted 2.56
HTTP Request Message Systems
Method URL HTTP Headers
version
Messag
e body
GET http://www.abc.ac.ind/data.html HTTP/1.1

Fig 2.12: HTTP Request Message

The following are the parts of HTTP request frame :
 Request Method: The request method indicates the method to be performed on
the resource identified by the given Request-URL. The method is case-
sensitive and should always be mentioned in uppercase.
 Request-URL: This Uniform Resource Identifier and identifies the resource
upon which to apply the request.
 Request Header Fields: The request-header fields allow the client to pass
additional information about the request, and about the client itself, to the
server. These fields act as request modifiers.
 HTTP Version: This specifies the version of HTTP used.
HTTP Methods
The following are the HTTP methods:
 GET: The GET method is used to retrieve information from the given server
using a given URI. Requests using GET should only retrieve data and should
have no other effect on the data.
 HEAD: Same as GET, but it transfers the status line and the header section
only.
 POST: A POST request is used to send data to the server, for example,
customer information, file upload, etc. using HTML forms.
 PUT: Replaces all the current representations of the target resource with the
uploaded content.
 DELETE: Removes all the current representations of the target resource given
by URI.
 CONNECT: Establishes a tunnel to the server identified by a given URI.
 OPTIONS: Describe the communication options for the target resource.
 TRACE: Performs a message loop back test along with the path to the target
resource.
 Distrubuted 2.57
Systems
HTTP Response Message
HTTP Version Status Code Reason Headers Message body
HTTP/1.1 2xx OK Data

Fig 2.13: HTTP Response Message

A response message consists of the protocol version followed by a numeric status
code and its associated textual phrase.
 HTTP Version: A server supporting HTTP version 1.1 will return the following
version information: HTTP-Version = HTTP/1.1
 Status Code: The Status-Code element is a 3-digit integer where first digit of the
Status-Code defines the class of response and the last two digits do not have any
categorization role. There are 5 values for the first digit:
 1xx- Informational :It means the request was received and the process is
continuing.
 2xx- Success: It means the action was successfully received, understood, and
accepted.
 3xx-Redirection: It means further action must be taken in order to complete
the request.
 4xx- Client Error: It means the request contains incorrect syntax or cannot be
fulfilled.
 5xx- Server error: It means the server failed to fulfill an apparently valid
request.
HTTP status codes are extensible and HTTP applications are not required to
understand the meaning of all registered status codes.
 Response Header Fields: The response-header fields allow the server to pass
additional information about the response which cannot be placed in the Status-
Line. These header fields give information about the server and about further
access to the resource identified by the Request-URI.
2.8.2 Remote Procedure Call (RPC)
RPC is a powerful technique for constructing distributed, client-server based
applications. In RPC, the procedure need not exist in the same address space as the calling
procedure. The two processes may be on the same system, or they may be on different
systems with a network connecting them. By using RPC, programmers of distributed
applications avoid the details of the interface with the network. RPC makes the client/server
model of computing more powerful and easier to program.
 Distrubuted 2.58
Systems
Remote Procedure Call (RPC) is a protocol that one program can use to request a
service from a program located in another computer in a network without having
to understand network details.

Design issues in RPC

The following are the design issues in RPC:

 Programming with interfaces

- Communication between modules in a program can be done by procedure calls
between modules or by direct access to the variables in another module.
- To control the interactions between modules, an explicit interface is defined for
each module that specifies the procedures and the variables that can be
accessed from other modules.
- The interface hides the programming details of one module from the other.
- The interaction between the modules can happen even after the modification of
the modules but without modifying the interfaces.

 Interfaces in distributed systems

- Each server in the client server model provides a set of procedures that are
available for use by clients.
- The separation between interface and implementation offers the following
advantages:
 The programmers need not be aware of implementation details of each
modules.
 In the perspective of distributed systems, the programmers need to know the
programming language or underlying platform used to implement the
service.
 It provides support for software evolution in which implementations can
change as long as long as the interface remains the same. But the interface
can also change as long as it remains compatible with the original.
 Interface definition languages:
- An RPC mechanism can be integrated with a particular programming language
if it contains notations for defining interfaces, allowing input and output
parameters to be mapped onto the language‟s normal use of parameters.
 Distrubuted 2.59
Systems
- But programs can also be written in a variety of languages to access them
remotely.
- In such cases Interface definition languages (IDL) are used which contains
procedures implemented indifferent languages to invoke one another.
 RPC call semantics
- In request-reply protocols doOperation can be implemented in different ways to
provide different delivery guarantees:
 Retry request message: Controls whether to retransmit the request message
until either a reply is received or the server is assumed to have failed.
 Duplicate filtering : Controls when retransmissions are used and whether to
filter out duplicate requests at the server.
 Retransmission of results: Controls whether to keep a history of result
messages to enable lost results to be retransmitted without re-executing the
operations at the server.
- The above choices lead to a variety of possible semantics as given below:
a) Maybe semantics: Here, the remote procedure call may be executed once
or not at all. Maybe semantics arises when no fault-tolerance measures are
applied and can suffer from the following types of failure:
 omission failures if the request or result message is lost
 crash failures.
May be semantics is useful only for applications in which occasional failed
calls are acceptable.
b) At-least-once semantics: Here the invoker receives either a result, in which
case the invoker knows that the procedure was executed at least once, or an
exception informing it that no result was received. At-least-once semantics
can suffer from the following types of failure:
 crash failures when the server containing the remote procedure fails
 arbitrary failures
c) At-most-once semantics: Here, the caller receives either a result, in which
case the caller knows that the procedure was executed exactly once, or an
exception informing it that no result was received, in which case the
procedure will have been executed either once or not at all.
 Transparency
- The remote procedure calls are much like local procedure calls as possible, with
no distinction in syntax between a local and a remote procedure call.
 Distrubuted 2.60
Systems
- All the necessary calls to marshalling and message-passing procedures were
hidden from the programmer making the call.
- Although request messages are retransmitted after a timeout, this is transparent
to the caller to make the semantics of remote procedure calls like that of local
procedure calls
- The usage of middleware can also offer additional levels of transparency to
RPC.
- But RPC‟s are more vulnerable to failure than local calls since they involve a
network, another computer and another process.
- So the clients making remote calls must be able to recover from such situations.
- The latency of a remote procedure call is several orders of magnitude greater
than that of a local one.
Implementation of RPC

Fig 2.14: Client and Server stub in RPC

- The client that accesses a service includes one stub procedure for each procedure
in the service interface.

A stub is a piece of code that is used to convert parameters during a remote

procedure call (RPC). An RPC allows a client computer to remotely call
procedures on a server computer.

- The parameters used in a function call have to be converted because the client and
server computers use different address spaces.

- Stubs perform this conversion so that the remote server computer perceives the
RPC as a local function call.

- The client stub routine performs the following functions:

 Distrubuted 2.61
Systems
 Marshals arguments: The client stub packages input arguments into a form
that can be transmitted to the server.
 Calls the client run-time library to transmit arguments to the remote address
space and invokes the remote procedure in the server address space.
 Unmarshals output argument: The client stub unpacks output arguments and
returns to the caller.
- The server stub procedure unmarshals the arguments in the request message, calls
the corresponding service procedure and marshals the return values for the reply
message.
- The client and server stub procedures and the dispatcher are automatically
generated by an interface compiler from the interface definition of the service.

2.8.3 Remote Method Invocation (RMI)

RMI is a set of protocols being developed by Sun's JavaSoft division that

enables Java objects to communicate remotely with other Java objects.

RMI is closely related to RPC but extended into the world of distributed objects. RMI
is a true distributed computing application interface specially designed for Java, written to
provide easy access to objects existing on remote virtual machines

Differences between RMI and RPC

RMI RPC
RMI uses an object oriented paradigm where RPC is not object oriented and does not deal
the user needs to know the object and the with objects. Rather, it calls specific
method of the object he needs to invoke. subroutines that are already established

With RPC looks like a local call. RPC RMI handles the complexities of passing
handles the complexities involved with along the invocation from the local to the
passing the call from the local to the remote remote computer. But instead of passing a
computer. procedural call, RMI passes a reference to the
object and the method that is being called.

The commonalities between RMI and RPC are as follows:

 They both support programming with interfaces.
 They are constructed on top of request-reply protocols.
 They both offer a similar level of transparency.
 unication in Distrubuted Distrubuted 2.62
2.62
Sytem Systems
Fig 2.15: RMI in JVM
Features of RMI
 In RMI, the remote objects are treated similarly to local objects.
 RMI provide access to objects existing on remote virtual machines.
 RMI handles marshalling, transportation, and garbage collection of the remote
objects.
 RMI was incorporated as a part of the JDK with version 1.1
Limitations of RMI
 RMI is not language independent. It is limited only to Java.
 Interfaces to remote objects are defined using ordinary Java interfaces not with
special IDLs.
Design issues of RMI
The RMI evolved from simple object model to distributed objects model.
a) Object Oriented Model
An object communicates with other objects by invoking their methods by
arguments and return values. Objects can encapsulate their dataand the code of
their methods. In a distributed object system, an object‟s data should be accessible
only through its methods.
 Object references:
– Objects are accessed through object references.
– Object references could be assigned to variables, passed as arguments and
returned as results of methods.
 Interfaces:
– An interface provides implementation independent, definition of the
signatures of a set of methods.
 Distrubuted 2.63
Systems
– An interface is not a class. A class describes the attributes and behaviors of
an object. An interface contains behaviors that a class implements.
– An interface can contain any number of methods.
– An interface is written in a file with a .java extension, with the name of the
interface matching the name of the file.
– The byte code of an interface appears in a .class file.
– Interfaces do not have constructors.
Differences between Interfaces and Class

Interfaces Class
Interfaces cannot be instantiated. Classes can be instantiated.
They do not contain constructors. They can have constructors.
They cannot have instance fields (i.e.) all the No constraints over the fields.
fields in an interface must be declared both
static and final.
The interface can contain only abstract They can contain method implementation.
method.
Interfaces are implemented by a class not Classes are extended by other classes.
extended.

 Actions :
– Action is initiated by an object invoking a method in another object with or
without arguments.
– The receiver executes the appropriate method and thenreturns control to the
invoking object.
– The following are the effects of actions:
 The state of the receiver may be changed.
 A new object may be instantiated.
 Further invocations on methods in other objects may take place.
 Exceptions:
– Exceptions are run time errors.
– The programmers must write code that could handle all possible unusual or
erroneous cases.
 Distrubuted 2.64
Systems
 Garbage collection:

– This is a method in Java to free the space occupied by obsolete objects.

b) Distributed objects

 Distributed object systems may adopt the client-server architecture where the
objects are managed by servers and their clients invoke their methods using
remote method invocation.

 In RMI, the client‟s request to invoke a method of an object is sent in a message

to the server managing the object.

 The invocation is carried out by executing a method of the object at the server
and the result is returned to the client in another message.

 Objects with different data formats may be used at different sites.

c) The distributed object model

In this model, each process contains a collection of objects, some of which can
receive both local and remote invocations, whereas the other objects can receive
only local invocations. The core concepts of distributed object model:

 Remote object references: Other objects can invoke the methods of a

remote object if they have access to its remote object reference.

 Remote interfaces: Every remote object has a remote interface that

specifies which of its methods can be invoked remotely.

 Remote object references: The notion of object reference is extended to

allow any object that can receive an RMI to have a remote object
reference.

A remote object reference is an identifier that can be used throughout a

distributed system to refer to a particular unique remote object.

The remote object to receive a remote method invocation is specified by the invoker as
a remote object reference. The remote object references may be passed as arguments and
results of remote method invocations.
 Remote interfaces: The class of a remote object implements the methods of its
remote interface. Objects in other processes can invoke only the methods that
belong to its remote
interface.
 Distrubuted 2.65
Systems

Fig 2.16: Remote and local invocations

Differences between remote and local objects

Remote Objects Local Objects

Creating a remote object requires creating a They consume less time cycles.
stub and a skeleton, which will cost more
time cycles.
Marshalling and unmarshalling is required. Marshalling and unmarshalling is not
required.
Longer time is taken by the remote object to Comparatively shorter return time.
return to the calling program.

Fig 2.17: Remote Object and remote

interfaces Garbage collection in a distributed-object system:
Distributed garbage collection is usually based on reference counting. The total
number of references to the object is maintained by a reference count field. The advantage
of using reference counts is that garbage is easily identified. When it becomes necessary to
reclaim the storage from unused objects, the garbage collector needs only to examine the
reference count fields of all the objects that have been created by the program. If the reference
count is zero, the object is garbage.
Exceptions:
Any remote invocation may fail because of exceptions. So RMI should include
exception handling as a part of their
implementation.
 Distrubuted 2.66
Systems
Implementation of RMI
Implementation of a RMI requires construction if many modules and the supporting
interfaces.
a) Modules:
The two common modules used are communication module and remote reference
module.
Communication module:
 This module implements request-reply protocol.
 This module specifies the message type, its requestId and the remote reference of
the object to be invoked.
 This specifies only the communication semantics.
 The communication module in the server selects the dispatcher for the class of the
object to be invoked, passing on its local reference, which it gets from the remote
reference module in return for the remote object identifier in the request message.
Remote reference module:
 This module is responsible for translation between local and remote object
references and for creating remote object references.
 This module has a remote object table specific for each process that records the
mapping between local object references and remote object references.
 The entries of the table includes:
 the remote objects held by the process.
 local proxy
 The actions of them remote reference module are as follows:
 When a remote object is to be passed as an argument for the first time, the
remote reference module creates a remote object reference and is added to the
table.
 When a remote object reference arrives in a request or reply message the
remote object table is referred for the corresponding local object reference.
 If the remote object reference is not in the table, the RMI software creates a
new proxy and asks the remote reference module to add it to the table.
 This module is called by components of the RMI software during marshalling
and unmarshalling remote object
references.
 Distrubuted 2.67
Systems
b) Supporting
Components Servants
 A servant is an instance of a class that provides the body of a remote object.
 The responsibility of the servant is to handle the remote requests passed on by the
corresponding skeleton.
 Servants live within a server process.
 They are created when remote objects are instantiated and remain in use until they
are no longer needed and finally servants are garbage collected or deleted.

The RMI software

 This consists of the application-level objects and the communication and remote
reference modules. The RMI software has the following components:
 Proxy: The class of the proxy contains a method corresponding to each
method in the remote interface which marshalls arguments and unmarshalls
results for that method. It hides the details of the remote object reference. Each
method of the proxy marshals a reference to the target object using a request
message (operationId, arguments). When the reply is received the proxy,
unmarshals it and returns the results to the invoker.
 Dispatcher: Server has a dispatcher & a skeleton for each class of remote
object. The dispatcher receives the incoming message, and uses its method info
to pass it to the right method in the skeleton.
 Skeleton: Skeleton implements methods of the remote interface to unmarshall
arguments and invoke the corresponding method in the servant. It then
marshalls the result (or any exceptions thrown) into a reply message to the
proxy.

Generation of the classes for proxies, dispatchers and skeletons

 The classes for the proxy, dispatcher and skeleton used in RMI are generated
automatically by an interface compiler.
 For Java RMI, the set of methods offered by a remote object is defined as a Java
interface.
Dynamic invocation:
 This is an alternative to proxies.
 The proxy just described (i.e.) its class is generated from an interface definition
and then compiled into the client code.
 Distrubuted 2.68
Systems
 If a client program receives a remote reference to an object whose remote interface
was not available at compile time. The method could not be invoked through
proxies.

 Dynamic invocation is used in this case which gives the client access to a generic
representation of a remote invocation, which is a part of the infrastructure for RMI.

 The client will supply the remote object reference, the name of the method and the
arguments to doOperation and then wait to receive the results.

Dynamic skeletons:

The Dynamic Skeleton Interface (or DSI) allows applications to provide

implementations of the operations on CORBA objects without static knowledge of the object's
interface. It is the server-side equivalent of the Dynamic Invocation Interface.

Server programs

The server program contains

 the classes for the dispatchers and skeletons, with the implementations of the
classes of all of the servants that it supports.

 an initialization section which is responsible for creating and initializing at

least one of the servants to be hosted by the server.

Client programs

The client program will contain

 the classes of the proxies for all of the remote objects that it will invoke.

Factory methods

 Factory methods are static methods that return an instance of the native class.

 Factory method is used to create different object from factory often refereed as
Item and it encapsulate the creation code.

 So instead of having object creation code on client side we encapsulate inside

Factory method in Java.

 The term factory method is sometimes used to refer to a method that creates
servants, and a factory object is an object with factory methods.

 Any remote object that needs to be able to create new remote objects on demand
for clients must provide methods in its remote interface for this purpose. Such
methods are called factory methods, although they are really just normal methods.
 Distrubuted 2.69
Binder Systems
 A binder in a distributed system is a separate service that maintains a table
containing mappings from textual names to remote object references.
 It is used by servers to register their remote objects by name and by clients to look
them up.
Server threads
 Whenever an object executes a remote invocation, that execution may lead to
further invocations of methods in other remote objects, which may take sometime
to return.
 To avoid this, servers generally allocate a separate thread for the execution of each
remote invocation.
Activation of remote objects
 Activation of servers to support remote objects is done by a service called Inetd.
 Certain applications demand the objects to persist for longer times.
 When the objects are maintained for longer period it is actually waste of resources.
 To avoid this the servers could be started on demand as and when needed through
Inetd.
 Processes that start server processes to host remote objects are called activators.
 Based on this remote objects can be described as active or passive.
 An object is said to be active when it is available for invocation.
 An object is called passive if is currently inactive but can be made active.
 A passive object consists of two parts:
 implementation of its methods
 its state in the marshalled form.
 When activating a passive object, an active object is created from the
corresponding passive object by creating a new instance of its class and initializing
its instance variables from the stored state.
 Passive objects can be activated on demand.
Responsibilities of an activator:

 Registering passive objects that are available for activation. This is done by
mapping the names of servers and passive objects.
 Distrubuted 2.70
Systems
 Starting named server processes and activating remote objects in them.
 Tracking the locations of the servers for remote objects that it has alreadyactivated.
 Java RMI provides the ability to make some remote objects activatable
[java.sun.comIX].

Persistent object stores

An object that is guaranteed to live between activations of processes is called a

persistent object.

 Persistent objects are managed by persistent object stores, which store their state in
a marshalled form on disk.
 The objects are activated when their methods are invoked by other objects.
 When activating the object, the invoker will not know whether an object is in main
memory or in a disk.
 The persistent objects that are not needed are maintained in the object store (active
or passive state).
 The decision of making an active object to passive is done through the following
ways:
 in response to a request in the program that activated the objects
 at the end of a transaction
 when the program exits.
 There are two approaches to determine whether an object is persistent or not:
 The persistent object store maintains some persistent roots, and any object that
is reachable from a persistent root is defined to be persistent.
 The persistent object store provides some classes. If the object belong to their
subclasses, then it is persistent.
Object location
 Remote object reference can be done based on the Internet address and port
number of the process that created the remote object.
 Some remote objects will exist in different processes, on different computers,
throughout their lifetime.
 In this case, a remote object reference cannot act as an address for the objects.
 Distrubuted 2.71
Systems
 So clients making invocations require both a remote object reference and an
address to which to send invocations.
 A location service helps clients to locate remote objects from their remote object
references.
 It acts as a database that maps remote object references to their probable current
locations.
 Alternatively a cache/broadcast scheme could also be used.
 Here, a member of a location service on each computer holds a small cache of
remote object reference to-location mappings.
 If a remote object reference is in the cache, that address is tried for the invocation
and will fail if the object has moved.
 To locate an object that has moved or whose location is not in the cache, the
system broadcasts a request.

Distributed garbage collection

 It is always desirable to automatically delete the remote objects that are no longer
referenced by any client.
 The memory held by those objects will be recovered.
 RMI uses a reference-counting garbage collection algorithm for automatic
garbage collection.
 In this algorithm, the RMI runtime keeps track of all live references within each
Java virtual machine.

 When a live reference enters a Java virtual machine, its reference count is
incremented.
 The first reference to an object sends a referenced message to the server for the
object.
 As live references are found to be unreferenced in the local virtual machine, the
count is decremented.
 When the last reference has been discarded, an unreferenced message is sent to the
server.
 The Java distributed garbage collection algorithm can tolerate the failure of client
processes.
 To achieve this, servers lease their objects to clients for a limited period of
time.
 Distrubuted 2.72
Systems
 The lease period starts when the client makes first referenceinvocation to the
server.
 It ends either when the time has expired or when the client makes a last reference
invocation to the server.
 The information stored by the server (i.e.) the leasecontains the identifier of the
client‟s virtual machine and the period of the lease.
 Clients are responsible for requesting the server to renew their leases before they
expire.

Leases in Jini
 Jini provides a mechanism for locating services on the network that conform to a
particular (Java) interface, or that have certain attributes associated with them.
 Once a service is located, the client can download an implementation of that
interface, which it then uses to communicate with the service.

 The Jini distributed system includes a specification for leases that can be used in a
variety of situations when one object offers a resource to another object.
 The granting of the use of a resource for a period of time is called a lease.
 The object offering the resource will maintain it until the time in the lease expires.
 The resource users are responsible for requesting their renewal when they expire.
 The period of a lease may be negotiated between the grantor and the recipient in
Jini.
 In Jini, an object representing a lease implements the Lease interface.
 It contains information about the period of the lease and methods enabling the
lease to be renewed or cancelled.

2.9 CASE STUDY: JAVA RMI

In this case study we use shared white board example. This program that allows a
group of users to share a common view of a drawing surface containing graphical objects,
such as rectangles, lines and circles, each of which has been drawn by one of the users.
Remote interfaces in Java RMI
 Remote interfaces are defined by extending an interface called Remote in the
java.rmipackage.
 The methods must throwRemoteException,contain an GraphicalObject which
implements theSerializableinterface.
 Distrubuted 2.73
Systems
Java Remote interfaces Shape and ShapeList
import java.rmi.*;
import java.util.Vector;
public interface Shape extends Remote
{
intgetVersion() throws RemoteException;
GraphicalObjectgetAllState() throws RemoteException;
}
public interface ShapeList extends Remote
{
Shape newShape(GraphicalObject g) throws RemoteException;
Vector allShapes() throws RemoteException;
intgetVersion() throws RemoteException;
}

Parameter and result passing

The parameters of a method are input parameters and the result of a method is a single
output parameter. Any object that implements the Serializable interface can be passed as an
argument or result in Java RMI. All primitive types and remote objects are serializable.
 Passing remote objects:
 When the type of a parameter or result value is defined as a remote
interface, the corresponding argument or result is always passed as a
remote object reference.
 When a remote object reference is received, it can be used to make
RMI calls on the remote object to which it refers.
 Passing non-remote objects:
 All serializable non-remote objects are copied and passed by value.
 When an object is passed by value, a new object is created in the
receiver‟s process.
 The methods of this new object can be invoked locally, possibly
causing its state to differ from the state of the original object in the
sender‟s process.
 In this example, the client uses the method new Shape that passes an
instance of GraphicalObject to the server.
 The server creates a remote object of type Shape with the state of the
GraphicalObject and returns a remote object reference to it.
 Distrubuted 2.74
Systems
 The arguments and return values in a remote invocation are used with
the following modifications:
1. Whenever an object that implements the Remote interface is
serialized, it is replaced by its remote object reference, which
contains the name of its (the remote object‟s) class.
2. When any object is serialized, its class information is annotated
with the location of the class (as a URL), enabling the class to be
downloaded by the receiver.

Downloading of classes
 In Java classes can be downloaded from one virtual machine to another.
 If the recipient does not already possess the class of an object passed by value, its
code is downloaded automatically.
 If the recipient of a remote object reference does not already possess the class for
a proxy, its code is downloaded automatically.
 The advantages are:
1. There is no need for every user to keep the same set of classes in their
working environment.
2. Both client and server programs can make transparent use of instances of
new classes whenever they are added.
RMIregistry
 The RMIregistry is the binder for Java RMI.
 It maintains a table mapping URL-style names to references to remote objects.
 The Naming class is used to access the RMIregistry.
 The methods of the naming class is of the general form:
//computerName:port/objectName
where computerName and port refer to the location of the RMIregistry.
 The LocateRegistryclass , which is in java.rmi.registry, is used to discover the
names.
 The getRegistry()returns an object of typeRegistry representing the remote binding
service:
public static Registry getRegistry() throws RemoteException
 Distrubuted 2.75
Systems
The following table gives a description of the methods in the Naming class:

Method Description
void rebind (String This method is used by a server to register the
name, Remote obj) identifier of a remote object by name.

void bind (String name, This method is also used to register a remote object by
Remote obj) name, but if the name is already bound to a remote
object reference an exception is thrown.
void unbind (String This method removes a binding.
name, Remote obj)
Remote lookup(String This method returns an array of Strings containing the
name)String [] list() names bound in the registry .This method is used by
clients to look up a remote object by name.

Server Program
import java.util.Vector;
public class ShapeListServant implements ShapeList
{
private Vector theList; // contains the list of Shapes
private int version;
public ShapeListServant(){...}
public Shape newShape(GraphicalObject g)
{
version++;
Shape s = new ShapeServant( g, version);
theList.addElement(s);
return s;
}
public Vector allShapes(){...}
public intgetVersion() { ... }
}
 Distrubuted 2.76
Systems
Client Program
import java.rmi.*;
import java.rmi.server.*;
import java.util.Vector;
public class ShapeListClient
{
public static void main(String args[])
{
System.setSecurityManager(new RMISecurityManager());
ShapeListaShapeList = null;
try
{
aShapeList = (ShapeList) Naming.lookup("//project.ShapeList");
Vector sList = aShapeList.allShapes();
}
catch(RemoteException e)
{
System.out.println(e.getMessage());
}
catch(Exception e)
{
System.out.println("Client: " + e.getMessage());
}
}
}

Callbacks
Callback refers to a server‟s action of notifying clients about an event. Callbacks can
be implemented in RMI as follows:
 The client creates a remote object that implements an interface that contains a
method for the server to call.
 The server provides an operation allowing interested clients to inform it of the
remote object references of their callback objects.
 Whenever an event of interest occurs, the server calls the interested clients.
 Distrubuted 2.77
Systems
The disadvantages of callbacks:
 The performance of the server may be degraded by the constant polling.
 Clients cannot notify users of updates in a timely manner.

Design and implementation of Java RMI

Use of reflection
 Reflection is used to pass information in request messages about the method to be
invoked.
 This is achieved with the help of the class Method in the reflection package.
 Each instance of Method represents the characteristics of a particular method,
including its class and the types of its arguments, return value and exceptions.
 An instance of Method can be invoked on an object of a suitable class by means of
its invoke method.
 The invoke method requires two arguments: the first specifies the object to receive
the invocation and the second is an array of Object containing the arguments. The
result is returned as type Object.
 The proxy has to marshal information about a method and its arguments into the
request message with array of Objects as its arguments and then marshals that
array.
 The dispatcher unmarshals the Method object and it s arguments in the array
of Objects fromethrequest message.
 The dispatcher then calls the Method object‟s invoke method, supplying the target
and the array of argument values. When the meth od has been executed, the
dispatcher marshals the result or any exceptions into the reply message.

Fig 2.13: Classes in Java RMI

Java classes supporting RMI
 Every servent need to extend UnicastRemoteObject.
 Distrubuted 2.78
Systems
 The class UnicastRemoteObject extends an abstract class called RemoteServer,
which provides abstract versions of the methods required by remote servers.
 Activatable class is available for providing activatable objects.
 The class RemoteServer is a subclass of RemoteObject that has an instance
variable holding the remote object reference and provides the following methods:
 equals: Compares remote object references.
 toString: Returns the contents of the remote object reference as a String
 readObject, writeObject: These methods deserialize/serialize remote objects

2.10 GROUP COMMUNICATION

Group communication is a service where a message is sent to a group and then
this message is delivered to all members of the group.

Programming Model
A group has group members (processes) who can join or leave the group. This is done
through multicast communication. There are three communication modes:
 Unicast: Process to process communication
 Broadcast: Communication to all the process
 Multicast: Communication only to a group of processes
Process Groups and Objects Groups
 An object group is a collection of objects that process the same set of invocations
concurrently.
 Client objects invoke operations on a single, local object, which acts as a proxy for
the group.
 The proxy uses a group communication system to send the invocations to the
members of the object group.
 Object parameters and results are marshalled as in RMI and the associated calls are
dispatched automatically to the right destination objects/methods.
Types of group communication
 Closed and open groups: A group is said to be closed if only members of the
group may multicast to it A process in a closed group delivers to itself any
message that it multicasts to the
group.
 Distrubuted 2.79
Systems

Fig 2.15: Open and Closed group communication

 Overlapping and non-overlapping groups: In overlapping groups, entities may
be members of multiple groups, and non-overlapping groups imply that
membership does not overlap to at most one.
 Synchronous and asynchronous systems: Group communication is possible in
both environments.
Reliability and ordering in multicast
 In group communication, all members of a group must receive copies of the
messages sent to the group, with delivery guarantees.
 Reliable multicast operation is based on:
 Integrity: delivering the message correctly only once
 Validity: guaranteeing that a message sent will eventually be delivered.
 Group communication demands relative ordering of messages delivered to
multiple destinations.
 The following are the types of message ordering:
 FIFO ordering: If a process sends one message before another, it will be
delivered in this order at all processes in the group.
 Causal ordering: If a message happens before another message in the
distributed
system this so-called causal relationship will be preserved in the delivery of
the associated messages at all processes.

 Total ordering: In total ordering, if a message is delivered before another

message at one process, then the same order will be preserved at all
processes.
 Distrubuted 2.80
Systems
Group membership management

A group membership service has four main tasks:

 Providing an interface for group membership changes: The membership

service provides operations to create and destroy process groups and to add or
withdraw a process to or from a group.

 Failure detection: The service monitors the group members in case of crash and
unreachability. The detector marks processes as Suspected or Unsuspected. The
process is excluded from membership if it is not reachable or crashed.

 Notifying members of group membership changes: The service notifies the

group‟s members when a process is added, or when a process is excluded.

 Performing group address expansion: When a process multicasts a message, it

supplies the group identifier rather than a list of processes in the group.

Fig 2.15: Group membership

management Case study: the JGroups toolkit

JGroups is a toolkit for reliable group communication written in Java. JGroups

supports process groups in which processes are able to join or leave a group, send a message
to all members of the group or indeed to a single member, and receive messages from the
group. The toolkit supports a variety of reliability and ordering guarantees.
 Distrubuted 2.81
Systems

Fig 2.16: Architecture of JGroups

The following are the important components in the architecture:
 Channels
 They represent the most primitive interface for application developers,
offering the core functions of joining, leaving, sending and receiving.
 A process interacts with a group through a channel object, which acts as a
handle onto a group.
 The following operations are allowed in channel:
 Connect: binds that handle to a particular named group.If the named
group does not exist, it is implicitly created at the time of the first
connect.
 Disconnect: To leave the group, the process executes the corresponding
disconnect operation.
 Close: will render the channel unusable.
 getView: returns the current view defined in terms of the current
member list
 getState: returns the historical application state associated with the
group
 Building blocks
 This offers higher-level abstractions, building on the underlying service
offered by channels.
 Distrubuted 2.82
Systems
 Some of the building blocks are:
 MessageDispatcher
- A sender to send a message to a group and then wait for some or
all of the replies.
- MessageDispatcher supports this by providing a castMessage
method that sends a message to a group and blocks until a
specified number of replies are received.
 RpcDispatcher
- This takes a specific method and invokes this method on all
objects associated with a group.
 NotificationBus
- This is an implementation of a distributed event bus, in which an
event is any serializable Java object.
- This class is often used to implement consistency in replicated
caches.
 Protocol stack
 This provides the underlying communication protocol, constructed as a
stack of protocol layers.
 Here, a protocol is a bidirectional stack of protocol layers with each
layer implementing the following two methods:
public Object up (Event evt);
public Object down (Event
evt);
 Each layer process the message, including modifying its contents, adding a
header or indeed dropping or reordering the message.
 This shows a protocol that consists of five layers:
- UDP: This is the transport layer that utilizes IP multicast and UDP
datagrams specifically for point-to-point communication. For larger-
scale systems operating over wide area networks, a TCP layer may be
preferred.
- FRAG: This implements message packetization and the maximum
packet size is 8,192 bytes by default.
- MERGE: This deal with unexpected network partitioning and the
subsequent merging of subgroups after the partition.
- GMS: This implements a group membership protocol to maintain
consistent views of membership across the group
- CAUSAL: This implements causal ordering.
 Distrubuted 2.83
Systems
2.11 PUBLISH -SUBSCRIBER SYSTEMS
 Publish–subscribe is a messaging pattern where senders of messages, called
publishers, do not program the messages to be sent directly to specific receivers,
called subscribers.
 Instead, published messages are characterized into classes. The subscribers express
interest in one or more classes, and only receive messages that are of interest.
 The task of the publish subscribe system is to match subscriptions against
published events and ensure the correct delivery of event notifications.
Characteristics of publish-subscribe systems
 Heterogeneity: In distributed environment, the different types of components
are made to interoperate with each other. To maintain integrity an interface is
designed for receiving and dealing with the resultant notifications.
 Asynchronicity: Notifications are sent asynchronously by event-generating
publishers to all the subscribers that have expressed an interest. This avoids the
need for synchronizing.
The programming model
 Publishers disseminate an event e through a publish(e) operation and subscribers
express an interest in a set of events through subscribe(f).
 This refers to a filter or a pattern defined over the set of all possible events.
 Subscribers can later revoke this interest through a corresponding unsubscribe(f)
operation.
 The events are delivered to the subscriber using a notify(e) operation.
 The advertise(f) allows the publishers to have the option of declaring the
nature of future events.

Fig 2.17: Publish Subscriber Model

 Distrubuted 2.84
Systems
The following are the subscriber or filter models:

 Channel-based:
 The publishers publish events to named channels and subscribers then
subscribe to one of these named channels to receive all events sent
to that channel.

 Topic-based (subject-based):
 Each notification from the publisher is expressed in terms of a number
of fields, with one field denoting the topic.
 Subscriptions are then defined in terms of the topic of interest.

 Content-based:
 Content-based approaches are a generalization of topic-based
approaches allowing the expression of subscriptions over a range of
fields in an event notification.

 Type-based:
 This is intrinsically linked with object-based approaches where objects
have a specified type.
 In type-based approaches, subscriptions are defined in terms of types of
events and matching is defined in terms of types or subtypes of the
given filter.

Implementation Issues
The publish subscribe model also demands of security, scalability, failure handling,
concurrency and quality of service. This makes this model complex.

Centralized versus distributed implementations

 The centralized implementation of this model is forming a single node with a
server on that node acting as an event broker.
 Publishers then publish events to this broker, and subscribers send subscriptions to
the broker and receive notifications in return.
 Interaction with the broker is then through a series of point-to-point messages; this
can be implemented using message passing or remote invocation.
 This lacks resilience and scalability, since the centralized broker represents a
single point for potential system failure and a performance bottleneck.
 Distrubuted 2.85
Systems

Fig 2.17: Network of Brokers

 Alternatively, the centralized broker is replaced by a network of brokers that
cooperate to offer the desired functionality.
 A fully peer-to-peer implementation of a publish-subscribe system is also possible.
 In this approach, there is no distinction between publishers, subscribers and
brokers; all nodes act as brokers, cooperatively implementing the required event
routing functionality.

Overall systems architecture

Fig 2.17: Overall system architecture

 Distrubuted 2.86
Systems
 In the bottom layer, publish-subscribe systems make use of a range of interprocess
communication services.

 The event routing layer is supported by a network overlay infrastructure. Event

routing performs the task of ensuring that event notifications are routed as
efficiently as possible to appropriate subscribers

 The overlay infrastructure supports this by setting up appropriate networks of

brokers or peer-to-peer structures.

 For content-based approaches, this problem is referred to as content-based

routing (CBR), with the goal being to exploit content information to efficiently
route events to their required destination.

 The top layer implements matching ensure that events match a given

subscription. The general principles behind content-based routing are:

 Flooding:

 Sending an event notification to all nodes in the network and then

carrying out the appropriate matching at the subscriber end is flooding.

 Alternatively, flooding can also be used to send subscriptions back to

all possible publishers, with the matching carried out at the publishing
end and matched events sent directly to the relevant subscribers using
point-to-point communication.

 Flooding can be implemented using an underlying broadcast or

multicast facility or arranging brokers in an acyclic graph in which each
forwards incoming event notifications to all its neighbors.

 Filtering:

 In filtering-based routing, filtering is done in the network of brokers.

 Brokers forward notifications through the network only where there is a

path to a valid subscriber.

 This is achieved by propagating subscription information through the

network towards potential publishers and then storing associated state
at each broker.

 Each node maintains a neighbors list containing a list of all connected

neighbors in the network of brokers, a subscription list containing a list
of all directly connected subscribers serviced by this node, and a
routing
table
 Distrubuted 2.87
Systems
 When a broker receives a publish request from a given node, it must
pass this notification to all connected nodes where there is a
corresponding matching subscription and also decide where to
propagate this event through the network of brokers.
 This approach can cause heavy traffic due to propagation of
subscription.
 Advertisements:
 Here the advertisements are propagated towards subscribers in a
symmetrical way.
 Rendezvous:
 Here, the set of all possible events are viewed as an event space and
responsibility is partitioned for this event space between the set of
brokers in the network.
 This approach defines rendezvous nodes (broker nodes), that are
responsible for a given subset of the event space.
 The rendezvous-based routing algorithm defines two functions.
 SN(s) takes a given subscription(s), and returns one or more
rendezvous nodes that take responsibility for that subscription. Each
rendezvous node maintains a subscription list and forwards all
matching events to the set of subscribing nodes.
 When an event e is published, the function EN(e) also returns one or
more rendezvous nodes, which are responsible for matching e
against subscriptions in the system.
 Both SN(s) and EN(e) return more than one node if reliability is a
concern.
 This approach works if the intersection of EN(e) and SN(s) is non-
empty for a given e that matches s .
Rendezvous-based routing algorithm
upon receive publish(event e) from node x at node i
rvlist := EN(e);
if i in rvlistthen begin
matchlist<- match(e, subscriptions);
send notify(e) to matchlist;
end
send publish(e) to rvlist - i;
 Distrubuted 2.88
Systems
upon receive subscribe(subscription s) from node x at node i
rvlist := SN(s);
if i in rvlistthen
add s to subscriptions;
else
send subscribe(s) to rvlist - i;

2.12 MESSAGE QUEUES

Message Queuing technology enables applications running at different times to
communicate across heterogeneous networks and systems that may be temporarily offline.
Applications send messages to queues and read messages from queues. Message Queuing
provides guaranteed message delivery, efficient routing, security, and priority-based
messaging.
It can be used to implement solutions to both asynchronous and synchronous scenarios
requiring high performance.
Programming Model
Communication in distributed systems is through queues. The producer processes can
send messages to a specific queue and consumer processes can receive messages from this
queue. There are three receiving styles
 blocking receive: block until an appropriate message is available
 non-blocking receive: check the status of the queue and return a message as
available, or not available.
 Notify operation: issue an event notification when a message is available in the
associated queue.
The queue can be build based on FIFO or priority.

Fig 2.17: Message Queue

 Distrubuted 2.89
Systems
Properties of message passing system:
 Reliability
 Persistence
 Validity
 Integrity
 Guaranteed delivery
 Atomicity
 Transformation messages between formats to deal with heterogeneity in
underlying data representations.
 Security

Implementation Issues
The important implementation decision for message queuing systems is the choice
between centralized and distributed implementations.
WebSphere MQ
 This is a IBM developed middleware based on the concept of message queues.
 Queues in WebSphere MQ are managed by queue managers which host and
manage queues and allow applications to access queues through the Message
Queue Interface(MQI).

 This allows applications to carry out operations such as connecting to or

disconnecting from a queue (MQCONN and MQDISC) or
sending/receiving messages to/from a queue (MQPUT and MQGET).
 Multiple queue managers can reside on a single physical server. Client
applications can also reside on the same physical server.
 The communication with the queue manager through is known as a client channel.
 MQI commands are issued on the proxy and then sent transparently to the queue
manager for execution using RPC.
 The message channel is a unidirectional connection between two queue managers
that is used to forward messages asynchronously from one queue to another.
 A message channel is managed by a message channel agent (MCA) at each end.
 Routing tables are also included in each queue manager, and together with
channels this allows arbitrary topologies to be created.
 Distrubuted 2.90
Systems

Fig 2.17: Network Topology in

Websphere The hub-and-spoke approach
 In the hub-and-spoke topology, one queue manager is designated as the hub.
 Client applications do not connect directly to this hub but rather connect through
queue managers called spokes.
 Spokes relay messages to the message queue of the hub for processing by the
various services.
 Spokes are placed strategically around the network to support different
clients.
 The hub is placed on a node with sufficient resources to deal with the volume of
traffic.
 This topology is heavily used with WebSphere MQ for large-scale deployments.
 The drawback of this architecture is that the hub can be a potential bottleneck and
a single point of failure.
2.12.1 Java Messaging Services (JMS)
 The Java Messaging Service (JMS) is a standardized way for distributed Java
programs to communicate indirectly.
 This unifies the publish-subscribe and message queue paradigms.
 A JMS client is a Java program or component that produces or consumes
messages.
 Distrubuted 2.91
Systems
Programming with JMS
 First create a connection between a client program and the provider thr ough a
connection factory.
 This is a logical channel between the client and provider.
 There are two types of connection can be established: Topic Connection or Queue
Connection.

 TopicConnection: support one or more topic sessions

 QueueConnection: support one or more queue sessions
 Connections can be used to create one or more sessions.
 A session is a series of operations involving the creation, productio n and
consumption of messages related to a logical task.

Fig 2.18: JMS Programming Model

Parts of a JMS Message

There are three parts: a header, a set of properties and the body of the message.

 Header
 The header contains all the information needed to identify and route the
message: destination, priority, expiration date, a message ID and a
timestamp.
 Distrubuted 2.92
Systems
 These fields are either created by the underlying system, or constructor
methods.
 The properties can be used to express additional context associated with the
message, including a location field.
 Body
 The body is opaque and untouched by the system.
 The body can be any one of a text message, a byte stream, a serialized
Java object, a stream of primitive Java values or a more structured set of
name/value pairs.
 A message producer is an object used to publish messages under a
particular topic or to send messages to a queue.
 A message consumer is an object used to subscribe to messages concerned
with a given topic or to receive messages from a queue.
 The consumer is complicated:
 Consumers apply message filters.
 Consumer program either can block using a receive operation or it
can establish a message listener object that identifies a message.

2.13 SHARED MEMORY APPROACHES

They are indirect communication paradigms that offer an abstraction of shared
memory.

Distributed shared memory (DSM)

Fig 2.18: Distributed Shared Memory

 Distrubuted 2.93
Systems
 This is an abstraction used for sharing data between computers that do not share
physical memory.
 Processes access DSM by reads and updates to what appears to be ordinary
memory within their address space.
 The programmer need not worry about the message passing.
 DSM runtime support has to send updates in messages between computers.
 DSM systems must also manage replicated data.
 Each computer has a local copy of recently accessed data items stored in DSM,
for speed of access.
Message passing versus DSM
The message passing and DSM can be compared based on services they offer and in
terms of their efficiency.

Message Passing Distributed Shared Memory

Services Offered:
Variables have to be marshalled The processes share variables directly, so
from one process, transmitted and no marshalling and unmarshalling. Shared
unmarshalled into other variables at the variables can be named, stored and
receiving process. accessed in DSM.

Processes can communicate with other Here, a process does not have private
processes. They can be protected from one address space. So one process can alter the
another by having private address spaces. execution of other.
This technique can be used in This cannot be used to heterogeneous
heterogeneous computers. computers.
Synchronization between processes is Synchronization is through locks and
through message passing primitives. semaphores.
Processes communicating via message Processes communicating through DSM
passing must execute at the same time. may execute with non-overlapping
lifetimes.
Efficiency:
All remote data accesses are explicit and Any particular read or update may or may
therefore the programmer is always aware not involve communication by the
of whether a particular operation is in- underlying runtime support.
process or involves the expense of
communication.
 Distrubuted 2.94
Systems
2.13.1 Tuple space communication
Processes communicate indirectly by placing tuples in a tuple space, from which other
processes can read or remove them.
The programming model
 Processes communicate through a shared collection of tuples.
 Any combination of types of tuples may exist in the same tuple space.
 Processes share data by accessing the same tuple space.
 The tuples are stored in tuple space using the write operation and read or extract
them from tuple space using the reador take operation.
 No direct access to tuples in tuple space is allowed and processes have to replace
tuples in the tuple space instead of modifying them. Thus, tuples are immutable.
Properties of tuple space
 Space uncoupling: A tuple placed in tuple space may originate from any number
of sender processes and may be delivered to any one of a number of potential
recipients.
 Time uncoupling: A tuple placed in tuple space will remain in that tuple space
until removed and hence the sender and receiver do not need to overlap in time.
Implementation Issues
Centralized solution where the tuple space resource is managed by a single server is a
common implementation.
Replication:
 A tuple space behaves like a state machine, maintaining state and changing this
state in response to events received from other replicas or from the environment.
 To ensure consistency the replicas
(ii) Must start in the same state
(iii) must execute events in the same order
(iv) must react deterministically to each event.
 Alternatively, a multicast algorithm can be used.
 Here, updates are carried out in the context of the agreed set of replicas and
tuples are also partitioned into distinct tuple sets based on their associated logical
names.
 The system consists of a set of workers carrying out computations on the tuple
space, and a set of tuple space
replicas.
 Distrubuted 2.95
 A given physical node can contain any number of workers,
Systems
replicas or indeed both;
a given worker therefore may or may not have a local replica.
 Nodes are connected by a communications network that may lose, duplicate or
delay messages and can deliver messages out of order.
 A write operation is implemented by sending a multicast message over the
unreliable communications channel to all members of the view.
 The write request is repeated until all acknowledgements are received.
 The read operation consists of sending a multicast message to all replicas.

Write
1. The requesting site multicasts the write request to all members of the view;
2. On receiving this request, members insert the tuple into their replica and
acknowledge this action;
3. Step 1 is repeated until all acknowledgements are received.

Read
1. The requesting site multicasts the read request to all members of the view;
2. On receiving this request, a member returns a matching tuple to the requestor;
3. The requestor returns the first matching tuple received as the result of the
operation (ignoring others);
4. Step 1 is repeated until at least one response is received.
Phase 1: Selecting the tuple to be removed
1. The requesting site multicasts the take request to all members of the view;
2. On receiving this request, each replica acquires a lock on the associated tuple set
and, if the lock cannot be acquired, the take request is rejected;
3. All accepting members reply with the set of all matching tuples;
4. Step 1 is repeated until all sites have accepted the request and responded with their
set of tuples and the intersection is non-null;
5. A particular tuple is selected as the result of the operation .
6. If only a minority accept the request, this minority are asked to release their locks
and phase 1 repeats.
Phase 2: Removing the selected tuple
1. The requesting site multicasts a remove request to all members of the view
citing the tuple to be removed;
 Distrubuted 2.96
Systems
2. On receiving this request, members remove the tuple from their replica, send
an acknowledgement and release the lock;
3. Step 1 is repeated until all acknowledgements are received

2.14 DISTRIBUTED OBJECTS

Distributed object communication realizes communication between distributed
objects in the distributed computing environment. The main role is to interconnect objects
residing in non-local memory space and allow them to perform remote calls and exchange
data. Distributed object middleware offers a programming abstraction based on object
oriented principles.
The following table gives a brief description about relation between objects and
distributed objects:

Objects Distributed Objects Descriptions

Object references Remote object Unique reference for each distributed object.
references
Interfaces Remote interfaces Provides methods to be invoked o remote
object using DLL.
Actions Distributed actions Invocation of methods
Exceptions Distributed Exception Includes exceptions generated in distributed
environment
Garbage Collection Distributed Garbage Implements distributed garbage collection
Collection algorithm

Differences between objects and distributed objects

Objects Distributed Objects

Class is a fundamental concept in object- Classes are not prominent here.
oriented languages.
In the object oriented world, class can be the Here the term „class‟ is avoided, more
description of the behavior associated with a specific terms such as „factory‟ and
group of objects or, the place to go to „template‟ are readily used.
instantiate an object with a given behavior or
even the group of objects that adhere
to that behavior.
Object oriented languages offer This offers interface inheritances.
implementation inheritances
 Distrubuted 2.97
Systems
Additional Complexities
 Inter-object communication:
 This is normally provided by remote method invocation.
 Lifecycle management:
 Lifecycle management is concerned with the creation, migration and deletion
of objects, with each step having to deal with the distributed nature of the
underlying environment.
 Activation and deactivation:
 In distributed systems, the numbers of objects may be very large, and hence it
would be waste of resources to have all objects available (active) at any time.
 Activation is the process of making an object active in the distributed
environment by providing the necessary resources for it to process incoming
invocations.
 Deactivation is rendering an object temporarily unable to process invocations.
 Persistence:
 Objects have state, and it maintains this state across possible cycles of
activation and deactivation and indeed system failures.
 Distributed object middleware must therefore offer persistency management
for stateful objects.
 Additional services:
 This must provide support for the range of distributed system services
considered in this book, including naming, security and transaction services.

2.15 FROM OBJECTS TO COMPONENTS

Component based computing is out powering distributed objects.

Problems in object-oriented middleware

The following four problems in object oriented middleware led to the development of
component based technology:

 Implicit dependencies:
 A distributed object communicates with outside world through interfaces.

 The interfaces provide a complete contract for the deploying and use of this
object.
 Distrubuted 2.98
Systems
 The problem in distributed objects is that the internal behavior of an
object is hidden.
 Implicit dependencies in the distributed configuration are not safe.
 The third-party developers to implement one particular element in a
distributed configuration.
Requirement: Dependencies of the object with otherobjects in the distributed
configuration must be specified.
 Interaction with the middleware:
 The distributed objects must interact even with low-levelmiddleware
architecture.
Requirement: While programming distributed applications, a clean separation
must be made between code related to middleware framework and code
associated with the application.
 Lack of separation of distribution concerns:
 Distributed programmers must also focus on issues related to security,
transactions, coordination and replication.
 To provide the non functional requirements :
 Programmers must have knowledge of the full details of all the
associated distributed system services.
 The implementation of an object will contain calls to distributed system
services and to middleware interfaces. This increases the programming
complexity.
Requirement: These complexities should be hidden from
the programmer.
 No support for deployment:
 Objects must be deployed manually on individual machines and activated
when required.
 This is a tiresome and error-prone process.
Requirement: Middleware platforms should support deployment.

2.15.1 Fundamentals of Components

A software component is a unit of composition with contractually specified

interfaces and explicit context dependencies only.
 Distrubuted 2.99
Systems
The term contract refers to:
 a set of provided interfaces
 a set of required interfaces
Every required interface must be bound to a provided interface of another
component. This is called software architecture. This consists of components, interfaces and
connections between interfaces.

Programming in component-based systems is concerned with the development of

components and their composition. So third party development of software components is
possible in component based technology (CBT).

Advantages of CBT
 Independent extensions
 Component Market
 Component models lessen unanticipated interactions between components
 Reduced time to market
 Reduced Costs

Disadvantages of CBT:
 Time to develop software components takes a big effort.
 Components can be pricey.

 Requirements in component technologies lacking

 Conflict between usability and reusability of components.
 Maintenance cost for components increased

Components and distributed systems Containers:

 The containers are absolutely central to component-based middleware.
 Containers support a common pattern often encountered in
distributed applications, which consists of:

o a front-end client
o a container holding one or more components that implement the
application or business logic
o system services that manage the associated data in persistent storage.
 Distrubuted 2.100
Systems
 The containers provide a managed server-side hosting environment
for components.
 The components deal with application concerns and the container deals with
distributed systems and middleware issues.
 A container includes a number of components that require the same configuration
in terms of distributed system support.
Support for Deployment
 Component-based middleware supports the deployment of
component configurations.
 They include releases of software architectures (components and
their interconnections) together with deployment descriptors.
 These descriptors describe how the configurations should be deployed in a
distributed environment.
 Deployment descriptors are typically written in XML and include sufficient
information to ensure that:
o components are correctly connected using appropriate protocols and
associated middleware support
o the underlying middleware and platform are configured to provide the right
level of support to the component configuration
o the associated distributed system services are set up to provide the right
level of security, transaction support and so on.

2.16 ENTERPRISE JAVA BEANS

 Enterprise JavaBeans (EJB) is a specification of a server-side, managed
component architecture
 It supports the development of the classic style of application, where
potentially large numbers of clients interact with a number of services realized
through components or configuration of components.
 The components, which are known as beans in EJB, are intended to capture the
application (or business) logic
 The container in EJB issues calls to the associated services to provide the
required properties.
 Beans act as:
 Container Managed: The transaction manager or security services is
completely hidden from the developer of the associated beans.
 Distrubuted 2.101
Systems
 Bean Managed: The bean developer takes control over these operations
not the transaction manager.
 The EJB specification has the following roles:
o bean provider: develops the application logic of the component(s)
o application assembler: assembles beans into application configurations
o deployer: takes a given application assembly and ensures it can be
correctly deployed in a given operational environment;
o service provider: does transaction management
o persistence provider: maps persistent data to databases
o container provider: build on the above two roles and is responsible for
configuring containers with the required level of distributed systems
support in terms of non-functional properties
o system administrator: responsible for monitoring a deployment at
runtime and making any adjustments to ensure its correct operation.
The EJB component model
 A bean in EJB is a component offering one or more business interfaces to
potential clients of that component.
 A bean is represented by the set of remote and local business interfaces
together with an associated bean class that implements the interfaces.
 Two main styles of bean are supported in the EJB specification:
 Session beans: A session bean is a component implementing a
particular task within the application logic of a service.
 Message-driven beans: Clients interact with session beans using local
or remote invocation.
POJOs and annotations
The task of programming in EJB has been simplified significantly through the use of
Enterprise JavaBean POJOs(plain old Java objects). It consists of the application
logic written simply in Java with no other code relating to it being a bean. Annotations
are then used to ensure the correct behaviour in the EJB context. Enterprise JavaBean
containers in EJB
 Beans are deployed to containers, and the containers provide implicit distributed
system management.
 The container provides the policies in areas including transaction management,
security, persistence and lifecycle management.
 Distrubuted 2.102
Systems
 The developer needs to focus only on application logic.
 EJB manages transactions.
 Transactions are sequences of operations, and that the sequences must be identified
by the transaction manager.
 The first thing to declare is whether transactions associated with an enterprise bean
should be bean-managed or container-managed. @TransactionManagement
(BEAN)@TransactionManagement (CONTAINER)
 The bean managed transaction uses the following methods:
 javax.transaction.UserTransaction– the User.Transaction.begin
 UserTransaction.commitmethods – within the code of the bean.

Transaction Attributes in EJB

Attribute Policy
REQUIRED If the client has an associated transaction running, execute
within this transaction; otherwise, start a new transaction.
REQUIRES_NEW Always start a new transaction for this invocation.
SUPPORTS If the client has an associated transaction, execute the method within
the context of this transaction; if not, the call proceeds without any
transaction support.
NOT_SUPPORTED If the client calls the method from within a transaction, then
this transaction is suspended before calling the method and
resumed afterwards – that is, the invoked method is excluded
from the transaction.
MANDATORY The associated method must be called from within a client
transaction; if not, an exception is thrown.
NEVER The associated methods must not be called from within a
client transaction; if this is attempted, an exception is
thrown

Dependency injection:
 In dependency injection a third party (container), is responsible for managing and
resolving the relationships between a component and its dependencies.
 A component refers to a dependency using an annotation and the container is
responsible for resolving this annotation and ensuring that, at runtime, the
associated attribute refers to the right object.
 This is typically implemented by the container usingreflection.
 Distrubuted 2.103
Systems
Enterprise JavaBean Interception
The Enterprise JavaBeans allows interception of two types of operation to alter their
default behavior:
• method calls associated with a business interface

• life cycle events.

Invocation Contexts

Method Description
public Object getTarget() Returns the bean instance associated with the incoming
invocation or event
public Method getMethod() Returns the method being invoked
public Object[] getParameters() Returns the set of parameters associated with the
intercepted business method

public void setParameters Allows the parameter set to be altered by theinterceptor,

(Object[] params) assuming type correctness ismaintained

public Object proceed() throws Execution proceeds to next interceptor in the chain
Exception (if any) or the method that has been intercepted
 Distrubuted 2.104
REVIEW QUESTIONS Systems

PART-A
1. What is IPC?
Interprocess communication (IPC) is a set of programming interfaces that allows a
programmer to coordinate activities among different program processes that can run
concurrently in an operating system.

2. List the models in distributed system.

 Physical models: It is the most explicit way in which to describe a system in terms
of hardware composition.

 Architectural models: They describe a system in terms of the computational and

communication tasks performed by its computational elements.

 Fundamental models: They examine individual aspects of a distributed system.

They are again classified based on some parameters as follows:

 Interaction models: This deals with the structure and sequencing of the
communication between the elements of the system

 Failure models: This deals with the ways in which a system may fail to operate
correctly

 Security models: This deals with the security measures implemented in the
system against attempts to interfere with its correct operation or to steal its
data.

3. Define ULS.

The ultra large scale (ULS) distributed systems is defined as a complex system
consisting of a series of subsystems that are systems in their own right and that come
together to perform a particular task or tasks.

4. What are the characteristics of ULS?

 very large size

 global geographical distribution

 operational and managerial independence of their member systems.

5. What are the evaluation elements of architectural model?

Architectural elements, architectural patterns and middleware platforms.

 Distrubuted 2.105
Systems
6. What are the architectural entities?
 Communicating entities (objects, components and web services);
 Communication paradigms (inter process communication, remote invocation and
indirect communication);
 Roles responsibilities and placement
7. What are the communicating entities in architectural model?
 Objects: They are used to implement object oriented approaches in distributed
systems. Objects are accessed through interfaces.
 Components: Components resemble objects and they offer problem-oriented
abstractions for building distributed systems. They are also accessed through
interfaces. The key difference between component and object is that a components
holds all the assumptions specified to other components and their interfaces in the
system. Components and objects are used to develop tightly coupled applications.
 Web services: Web services are closelyrelated to objects and components. Web
services areintegrated into the World Wide Web. They are partially defined by the
web-based technologies theyadopt. Web services are generally viewed as complete
servicesthat can be combined to achieve value-added services across
organizational boundaries.
8. List the communication mechanisms in DS.
IPC, Remote invocation and indirect communication.
9. Give the features of IPC.
This has the following two key features:
 Space uncoupling: Senders do not need to know who they are sending to
 Time uncoupling: Senders and receivers do not need to exist at the same time
10. What is Group communication?
Group communication is concerned with the delivery of messages to a set of recipients
and hence is a multiparty communication paradigm supporting one-to-many
communication. A group identifier uniquely identifies the group. The recipients join
the group and receive the messages. Senders send messages to the group based on the
group identifier and hence do not need to know the recipients of the message.
11. What are Publish-subscribe systems?
This is a messaging pattern where senders of messages, called publishers, do not
program the messages to be sent directly to specific receivers, called subscribers.
Instead, published messages are characterized into classes, without knowledge of
what, subscribers there may be. Similarly, subscribers express interest in one or more
classes, and only receive messages that are of interest, without knowledge of what, if
any, publishers there are. They offer one-to-many style of communication.
 Distrubuted 2.106
Systems
12. What are Message queues?
They offer a point-to-point service whereby producer processes can send messages to
a specified queue and consumer processes can receive messages from the queue or be
notified of the arrival of new messages in the queue. Queues therefore offer an
indirection between the producer and consumer processes.
13. What are Tuple spaces?
The processes can place arbitrary items of structured data, called tuples, in a persistent
tuple space and other processes can either read or remove such tuples from the tuple
space by specifying patterns of interest. This style of programming is known as
generative communication.
14. Write about Distributed shared memory.
In computer architecture, distributed shared memory (DSM) is a form of memory
architecture where the (physically separate) memories can be addressed as one
(logically shared) address space. Here, the term shared does not mean that there is a
single centralized memory but shared essentially means that the address space is
shared (same physical address on two processors refers to the same location in
memory
15. What are the architectures based on roles and responsibilities?
Client server and peer to peer.
16. What are the strategies for placement of objects?
 Mapping of services to multiple servers
 Caching
 Mobile code
 Mobile agents
17. What is mobile code?
Applets are a well-known and widely used example of mobile code. The user running
a browser selects a link to an applet whose code is stored on a web server; the code is
downloaded to the browser and runs. They provide interactive response.
18. What are mobile agents?
A mobile agent is a running program (including both code and data) that travels from
one computer to another in a network carrying out a task on someone‟s behalf, such as
collecting information, and eventually returning with the results. A mobile agent is a
complete program, code + data, that can work (relatively) independently.
19. What is layering?
In a layered approach, a complex system is partitioned into a number of layers, with a
given layer making use of the services offered by the layer below.
 Distrubuted 2.107
Systems
20. Define tiering.
Tiering is a technique to organize functionality of a given layer and place this
functionality into appropriate servers and, as a secondary consideration, on to physical
nodes.

21. Write about two tier architecture.

The two tiered architecture refers to client/server architectures in which the user
interface (presentation layer) runs on the client and the database (data layer) is stored
on the server. The actual application logic can run on either the client or the server.

22. Write about three tier architecture.

Presentation tier: This is the topmost level of the application. is concerned with
handling user interaction and updating the view of the application as presented to the
user;
Application tier (business logic, logic tier, or middle tier): The logical tier is pulled out
from the presentation tier and, as its own layer, it controls an application‟s
functionality by performing detailed processing.
Data tier: The data tier includes the data persistence mechanisms (database servers, file
shares, etc.) and the data access layer that encapsulates the persistence mechanisms
and exposes the data.

23. Define thin client.

In a thin-client model, all of the application processing and data management is carried
out on the server. The client is simply responsible for running the responsible for
running the presentation software. This is used when legacy systems are migrated to
client server architectures.

24. What is VNC?

The virtual network computing (VNC) has emerged to overcome the disadvantages of
thin clients. VNC is a type of remote-control software that makes it possible to control
another computer over a network connection.

25. What is Proxy?

 This facilitates location transparency in remote procedure calls or remote method
invocation.
 A proxy is created in the local address space to represent the remote object with
same interface as the remote object.
 The programmer makes calls on this proxy object and he need not be aware of the
distributed nature of the interaction.
 Distrubuted 2.108
Systems
26. Define Brokerage.
 It is used to bring interoperability in potentially complex
distributed infrastructures.
 The service broker is meant to be a registry of services, and stores information
about what services are available and who may use them.
27. What is reflection?
 The Reflection architectural pattern provides a mechanism for changing structure
and behaviour of software systems dynamically.
 It supports the modification of fundamental aspects, such as type structures and
function call mechanisms.
28. What are the levels in reflection?
Meta Level: This provides information about selected system properties and makes the
software self-aware.
Base level: This includes the application logic. Its implementation builds on the meta
level. Changes to information kept in the meta level affect subsequent base-level
behavior.
29. Define middleware.
Middleware is a general term for software that serves to "glue together" separate, often
complex and already existing, programs.
30. What are Distributed Objects?
The term distributed objects usually refers to software modules that are designed to
work together, but reside either in multiple computers connected via a network or in
different processes inside the same computer. One object sends a message to another
object in a remote machine or process to perform some task. The results are sent back
to the calling object.
31. What are Distributed Components?
A component is a reusable program building block that can be combined with other
components in the same or other computers in a distributed network to form an
application. Examples: a single button in a graphical user interface, a small interest
calculator, an interface to a database manager. Components can be deployed on
different servers in a network and communicate with each other for needed services. A
component runs within a context called a container . Examples: pages on a Web site,
Web browsers, and word processors.
32. Write about Publish subscriber model.
Publish–subscribe is a messaging pattern where senders of messages, called
publishers, do not program the messages to be sent directly to specific receivers, called
 Distrubuted 2.109
Systems
subscribers. Instead, published messages are characterized into classes, without
knowledge of what, if any, subscribers there may be. Similarly, subscribers express
interest in one or more classes, and only receive messages that are of interest, without
knowledge of what publishers are there.
33. Write about Message Queues.
Message queues provide an asynchronous communications protocol, meaning that the
sender and receiver of the message do not need to interact with the message queue at
the same time. Messages placed onto the queue are stored until the recipient retrieves
them. Message queues have implicit or explicit limits on the size of data that may be
transmitted in a single message and the number of messages that may remain
outstanding on the queue.

34. Define Web services.

A web service is a collection of open protocols and standards used for exchanging
data between applications or systems. Software applications written in various
programming languages and running on various platforms can use web services to
exchange data over computer networks like the Internet in a manner similar to inter-
process communication on a single computer. This interoperability (e.g., between Java
and Python, or Windows and Linux applications) is due to the use of open standards.

35. What is Peer to peer computing?

Peer-to-peer (P2P) computing or networking is a distributed application architecture
that partitions tasks or work load between peers. Peers are equally privileged,
equipotent participants in the application. They are said to form a peer-to-peer network
of nodes.

36. List the advantages of middleware

 Real time information access among systems.
 Streamlines business processes and helps raise organizational efficiency.
 Maintains information integrity across multiple systems.
 It covers a wide range of software systems, including distributed Objects and
components, message-oriented communication, and mobile application support.
 Middleware is anything that helps developers create networked applications.

37. List the disadvantages of middleware

 Prohibitively high development costs.
 There are only few people with experience in the market place to develop and use
a middleware.
 Distrubuted 2.110
Systems
 There are only few satisfying standards.
 The tools are not good enough.
 Too many platforms to be covered.
 Middleware often threatens the real-time performance of a system.

 Middleware products are not very mature.

38. Define fundamental models.

Fundamental Models are concerned with a formal description of the properties that are
common in all of the architectural models.

39. What are the types of fundamental models?

Interaction Model – deals with performance and the difficulty of setting of time limits
in a distributed system.
Failure Model – specification of the faults that can be exhibited by processes
Secure Model – discusses possible threats to processes and communication channels.

40. Give the purpose of the fundamental model.

 Make explicit all the relevant assumptions about the systems we are modelling.
 Make generalizations (general purpose algorithms) concerning what is possible or
impossible with given assumptions.
41. What are the communication performance metrics?
a. Latency: A delay between the start of a message‟s transmission from one process
to the beginning of reception by another.
b. Bandwidth: The total amount of information that can be transmitted over in a given
time. The communication channels using the same network, have to share the
available bandwidth.
c. Jitter: The variation in the time taken to deliver a series of messages. It is very
relevant to multimedia data.

42. What is synchronous DS?

 This is practically hard to achieve in real life.
 In synchronous DS the time taken to execute a step of a process has known lower
and upper
bounds.
 Distrubuted 2.111
Systems
 Each message transmitted over a channel is received within a known bounded
time.
 Each process has a local clock whose drift rate from real time has known bound.

43. What is Asynchronous DS?

 There are no bounds on: process execution speeds, message transmission delays
and clock drift rates.

44. What are the common types of failures?

 Omission Failure
 Arbitrary Failure
 Timing Failure

45. What is Omission failure?

Omission failures occur due to communication link failures. They are detected through
timeouts.
46. Write about Process omission failures.
 Omission failures that occur due to process crash (i.e.) the execution of the process
could not continue.

 This is detected using timeouts.

 A fixed period of time is fixed for all the methods to complete its execution. If the
method takes time longer than the allowed time, a time out has occurred.
 In an asynchronous system a timeout can indicate only that a process is not
responding.
 A process crash is called fail-stopif other processes can detect certainly that the
process has crashed.
 Fail-stop behavior can be produced in a synchronous system if the processes use
timeouts to detect when other processes fail to respond and messages are
guaranteed to be delivered.

47. Write about Communication omission failures.

 This occurs when the messages are dropped between sender and the receiver.
 The message can be lost in sender buffer, receiver buffer or even in the
communication
channel.
 Distrubuted 2.112
Systems
48. What is send omission failure?
The loss of messages between the sending process and the outgoing message buffer is
known as send omission failures.
49. What is receive-omission failure?
The loss of messages between the incoming message buffer and the receiving process
as receive-omission failures.

50. What is channel omission failure?

The loss of messages in a channel is channel omission failure.

51. What are Arbitrary failures (Byzantine failure)?

This includes all possible errors that could cause failure.

52. Classify the arbitrary failures.

Class Affects Comments

Fail stop Process Process halts and remains halted. Other
processes may detect that this process has
halted.
Crash Process Process halts and remains halted. Other
processes may not detect that this process
has halted.
Omission Channel A message inserted in an outgoing
message buffer never arrives at the other
end‟s incoming messagebuffer.
Send-omission Process A process completes a send operation but
the message is not put in its outgoing
message buffer.
Receiveomission Process A message is put in a process‟s incoming
messagebuffer, but that process does not
receive it.

Arbitrary Process Process/channel exhibits arbitrary

or behaviour: it may send/transmit arbitrary
messages at arbitrary times or commit
channel omissions; a process may stop or take an
incorrect step.
 Distrubuted 2.113
Systems
53. Define timing failure.
Timing failures refer to a situation where the environment in which a system operates
does not behave as expected regarding the timing assumptions, that is,
the timing constraints are not met.
54. Define reliable communication.
The term reliable communication is defined in terms of validity and integrity.
 Validity: Any message in the outgoing message buffer is eventually delivered to
the incoming message buffer.
 Integrity: The message received is identical to one sent, and no messages are
delivered twice.
55. What is security model?
The security of a DS can be achieved by securing the processes and the channels used
in their interactions and by protecting the objects that they encapsulate against
unauthorized access.
56. What is DOS attack?
Denial of service (DoS) attack is an incident in which a user or organization is
deprived of the services of a resource they would normally expect to have. In a
distributed denial-of-service, large numbers of compromised systems (sometimes
called a botnet) attack a single target.
57. Define IPC.
Interprocess communication (IPC) is a set of programming interfaces that allow a
programmer to coordinate activities among different program processes that can run
concurrently in an operating system.
58. What are the characteristics of IPC?
Characteristics of IPC:
 Synchronous and asynchronous communication
 Message destinations
 Reliability
 Ordering
59. Define socket.
A socket is one endpoint of a two-way communication link between two programs
unning on the
network.
 Distrubuted 2.114
Systems
60. What are the UDP failure modes?
UDP datagrams suffer from following failures:
 Omission failure: Messages may be dropped occasionally
 Ordering: Messages can be delivered out of order.
61. List the issues in stream communication.
Matching of data items: Two communicating processes need to agree as to the
contents of the data transmitted over a stream.
Blocking: The process that writes data to a stream may be blocked by the TCP flow-
control mechanism if the socket at the other end is queuing as much data as the
protocol allows.
Threads: When a server accepts a connection, it generally creates a new thread in
which to communicate with the new client. In an environment in which threads are not
provided, an alternative is to test whether input is available from a stream before
attempting to read it;
62. Define external data representation.
External Data Representation is an agreed standard for the representation of data
structures and primitive values.
63. Define marshalling and unmarshalling.
Marshalling is the process of taking a collection of data items and assembling them
into a form suitable for transmission in a message. Unmarshalling is the process of
disassembling a collection of data on arrival to produce an equivalent collection of
data items at the destination.
64. Define reflection.
Reflection is inquiring about class properties, e.g., names, types of methods and
variables, of objects
65. List the requirements of remote object reference.
The following are mandatory for remote object reference
 internet address/port number: process which created object
 time: creation time
 object number: local counter, incremented each time an object is created in the
creating process
 interface: how to access the remote object (if object reference is passed from one
client to another)
 Distrubuted 2.115
Systems
66. Define multi cast communication.
Multicast (one-to-many or many-to-many distribution) is group communication where
information is addressed to a group of destination computers simultaneously.
67. List the characteristics of multicasting.
The following are the characteristics of multicasting:
 Fault tolerance based on replicated services:
 Finding the discovery servers in spontaneous networking
 Better performance through replicated data
 Propagation of event notifications
68. What is IP multicast?
IP multicast is built on top of the Internet protocol. IP multicast allows the sender to
transmit a single IP packet to a multicast group. A multicast group is specified by class
D IP address with1110 as it starting byte.
69. List some methods in MulticastSocket class.
 getInterface():Retrieve the address of the network interface used for multicast
packets.
 getTTL():Get the default time-to-live for multicast packets sent out on the socket.
 joinGroup(InetAddress): Joins a multicast group.
 leaveGroup(InetAddress):Leave a multicast group.
 send(DatagramPacket, byte):Sends a datagram packet to the destination, with a
TTL (time- to-live) other than the default for the socket.
 setInterface(InetAddress): Set the outgoing network interface for multicast packets
on this socket, to other than the system default.
 setTTL(byte): Set the default time-to-live for multicast packets sent out on this
socket.
70. Define network virtualization.
Network virtualization refers to the management and monitoring of an entire computer
network as a single administrative entity from a single software-based administrator‟s
console.
71. Define overlay network.
An overlay network is a virtual network of nodes and logical links that is built on top
of an existing network with the purpose to implement a network service that is not
available in the existing
network.
 Distrubuted 2.116
Systems
72. List the Advantages of overlay network.
 They enable new network services to be defined without requiring changes to the
underlying network.
 They encourage experimentation with network services and the customization of
services to particular classes of application.
 Multiple overlays can be defined and can coexist, with the end result being a more
open and extensible network architecture.
73. List the disadvantages of overlay network.
 They introduce an extra level of indirection.
 They add to the complexity of network services
74. Define message passing interface.
Message Passing Interface (MPI) is a standardized and portable message-passing
system to function on a wide variety of parallel computers.
75. What are the types of message passing?
 Synchronous message passing: This is implemented by blocking send and receive
calls.
 Asynchronous message passing: This is implemented by a non-blocking form of
send.
76. What is remote invocation?
Remote invocations are method invocations for objects in different processes.
77. What are remote objects?
Remote objects are objects that can receive remote invocation. The core part of the
distributed object model is:
- remote object reference;
- remote interface: specifies which methods can be invoked remotely;
78. Give the advantages of UDP over TCP.
UDP is preferred over TCP for the following reasons:
 Acknowledgements are redundant, since requests are followed by replies.
 Establishing a connection involves two extra pairs of messages in addition to the
pair required for a request and a reply.
 Flow control is redundant for the majority of invocations, which pass only small
arguments and results.
 Distrubuted 2.117
Systems
79. What is message id?
Each message have a unique message identifier for referencing it. A message
identifier consists of two parts:
 request Id: Increasing sequence of integers assigned by the sending process. This
makes the message unique to the sender.
 an identifier: for the sender process. This makes the message unique in the
distributed system.
80. What is history of messages?
History is used to refer to a structure that contains a record of reply messages that have
been transmitted by the server.
81. List the styles of exchange protocols.
There are three exchange protocols:
 request (R) protocol: a single request message is sent by the client to the server
 request-reply (RR) protocol: used in client-server exchanges because it is based on
the request-reply protocol. Special acknowledgement messages are not required.
 request-reply-acknowledge reply (RRA) protocol: It is based on request-reply-
acknowledge reply. The Acknowledge reply message contains the requested from
the reply message being acknowledged. This will enable the server to discard
entries from its history. The arrival of a requested in an acknowledgement message
will be interpreted as acknowledging the receipt of all reply messages with lower
request Ids.
82. Give the differences between persistent and non persistent connections

Persistent Connection Non persistent Connection

On the same TCP connection the The server parses request, responds, and
server, parses request, responds and closes TCP connection.
waits for new requests.
It takes fewer RTTs to fetch the It takes 2 RTTs to fetch each object.
objects.
It has less slow start. Each object transfer suffers from slow start

83. Define RPC.

Remote Procedure Call (RPC) is a protocol that one program can use to request a
service from a program located in another computer in a network without having to
understand network
details.
 Distrubuted 2.118
Systems
84. List the design issues of RPC.
The following are the design issues in RPC:
 Programming with interfaces
 Interfaces in distributed systems
 Interface definition languages:
 RPC call semantics

85. Define stub.

A stub is a piece of code that is used to convert parameters during a remote procedure
call (RPC). An RPC allows a client computer to remotely call procedures on a server
computer.

86. Define RMI.

RMI is a set of protocols being developed by Sun's JavaSoft division that
enables Java objects to communicate remotely with other Java objects.

87. List the differences between RMI and RPC

RMI RPC

RMI uses an object oriented paradigm RPC is not object oriented and does not
where the user needs to know the object deal with objects. Rather, it calls specific
and the method of the object he needs subroutines that are already established
to invoke.

With RPC looks like a local call. RPC RMI handles the complexities of passing
handles the complexities involved with along the invocation from the local to the
passing the call from the local to the remote computer. But instead of passing a
remote computer. procedural call, RMI passes a reference to
the object and the method that is being
called.

88. Give the Limitations of RMI

 RMI is not language independent. It is limited only to Java.
 Interfaces to remote objects are defined using ordinary Java interfaces not with special
IDLs.
 Distrubuted 2.119
Systems
89. Give the differences between Interfaces and Class

Interfaces Class
Interfaces cannot be instantiated. Classes can be instantiated.
They do not contain constructors. They can have constructors.
They cannot have instance fields (i.e.) all No constraints over the fields.
the fields in an interface must be declared
both static and final.
The interface can contain only abstract They can contain method
method. implementation.
Interfaces are implemented by a class not Classes are extended by other
extended. classes.

90. What is remote object reference?

A remote object reference is an identifier that can be used throughout a distributed
system to refer to a particular unique remote object.

91. Give the differences between remote and local objects

Remote Objects Local Objects

Creating a remote object requires creating a They consume less time cycles.
stub and a skeleton, which will cost more
time cycles.
Marshalling and unmarshalling is required. Marshalling and unmarshalling is not
required.
Longer time is taken by the remote object to Comparatively shorter return time.
return to the calling program.

92. What is reference counting?

Distributed garbage collection is usually based on reference counting. The total
number of references to the object is maintained by a reference count field.
93. Give the use of Proxy in RMI.
The class of the proxy contains a method corresponding to each method in the remote
interface which marshalls arguments and unmarshalls results for that method. It hides
the details of the remote object reference. Each method of the proxy marshals a
reference to the target object using a request message (operationId, arguments). When
the reply is received the proxy, unmarshals it and returns the results to the invoker.
 Distrubuted 2.120
Systems
94. Give the use of Dispatcher in RMI.
Server has a dispatcher & a skeleton for each class of remote object. The dispatcher
receives the incoming message, and uses its method info to pass it to the right method
in the skeleton.
95. Give the use of Skeleton in RMI.
Skeleton implements methods of the remote interface to unmarshall arguments and
invoke the corresponding method in the servant. It then marshalls the result (or any
exceptions thrown) into a reply message to the proxy.
96. What are factory methods?
Factory methods are static methods that return an instance of the native class.Factory
method is used to create different object from factory often refereed as Item and it
encapsulate the creation code.
97. List the responsibilities of
activator. Responsibilities of an
activator:
 Registering passive objects that are available for activation. This is done by mapping
the names of servers and passive objects.
 Starting named server processes and activating remote objects in them.
 Tracking the locations of the servers for remote objects that it has alreadyactivated.
 Java RMI provides the ability to make some remote objects
activatable [java.sun.comIX].
98. Define persistent object.
An object that is guaranteed to live between activations of processes is called a
persistent object.
99. What is Jini?
Jini provides a mechanism for locating services on the network that conform to a
particular (Java) interface, or that have certain attributes associated with them. Once a
service is located, the client can download an implementation of that interface, which
it then uses to communicate with the service.
100. What is callback?
Callbackrefers to a server‟s action of notifying clients about an event.
101. Define group communication.
Group communication is a service where a message is sent to a group and then this
message is delivered to all members of the group.
 Distrubuted 2.121
Systems
102. What is publish subscribe model?
Publish–subscribe is a messaging pattern where senders of messages, called
publishers, do not program the messages to be sent directly to specific receivers,
called subscribers.
103. What is message queuing?
Message Queuing technology enables applications running at different times to
communicate across heterogeneous networks and systems that may be temporarily
offline.
104. Define session.
A session is a series of operations involving the creation, production and
consumption of messages related to a logical task.
105. What are the parts of JMS message?
There are three parts: a header, a set of properties and the body of the message.
106. What is DSM?
This is an abstraction used for sharing data between computers that do not share
physical memory. Processes access DSM by reads and updates to what appears to be
ordinary memory within their address space.
107. Give the differences between message passing and DSM.

Message Passing Distributed Shared Memory

Services Offered:
Variables have to be marshalled The processes share variables directly, so
from one process, transmitted and no marshalling and unmarshalling. Shared
unmarshalled into other variables at the variables can be named, stored and accessed
receiving process. in DSM.
Processes can communicate with other Here, a process does not have private
processes. They can be protected from one address space. So one process can alter the
another by having private address spaces. execution of other.
This technique can be used in This cannot be used to heterogeneous
heterogeneous computers. computers.
Synchronization between processes is Synchronization is through locks and
through message passing primitives. semaphores.
Processes communicating via message Processes communicating through DSM
passing must execute at the same time. may execute with non-overlapping
lifetimes.
 Distrubuted 2.122
Efficiency: Systems
All remote data accesses are explicit and Any particular read or update may or may
therefore the programmer is always aware not involve communication by the
of whether a particular operation is in- underlying runtime support.
process or involves the expense of
communication.

108. Give the properties of tuple space.

 Space uncoupling: A tuple placed in tuple space may originate from any number
of sender processes and may be delivered to any one of a number of potential
recipients.
 Time uncoupling: A tuple placed in tuple space will remain in that tuple space
until removed and hence the sender and receiver do not need to overlap in time.

109. Give the differences between objects and distributed objects.

Objects Distributed Objects

Class is a fundamental concept in object- Classes are not prominent here.
oriented languages.
In the object oriented world, class can be Here the term „class‟ is avoided, more
the description of the behavior associated specific terms such as „factory‟ and
with a group of objects or, the place to go „template‟ are readily used.
to instantiate an object with a given
behavior or even the group of objects that
adhere
to that behavior.
Object oriented languages offer This offers interface inheritances.
implementation inheritances

110. Define software component.

A software component is a unit of composition with contractually specified interfaces
and explicit context dependencies only.

111. Give the advantages and disadvantages of CBT.

Advantages of CBT

 Independent extensions
 Distrubuted 2.123
Systems
 Component Market
 Component models lessen unanticipated interactions between components
 Reduced time to market
 Reduced Costs
Disadvantages of
CBT:
 Time to develop software components takes a big effort.

 Components can be pricey.

 Requirements in component technologies lacking
 Conflict between usability and reusability of components.
 Maintenance cost for components increased

112. What is EJB?

Enterprise JavaBeans (EJB) is a specification of a server-side, managed component
architecture. It supports the development of the classic style of application, where
potentially large numbers of clients interact with a number of services realized
through components or configuration of components.

PART-B
1. Describe the physical system model.
2. Explain architectural models.
3. Write notes about architectural patterns.
4. Explain the middleware.
5. Describe the fundamental models.
6. Explain about interprocess communication.
7. Describe external data representation and marshaling.
8. Elucidate multicast communication.
9. Explain overlay networks.
10. Write notes on message passing interface.

11. Describe remote invocation.

12. Write in detail about Java RMI.
 Distrubuted 2.124
Systems
13. Explain group communications.
14. Describe publish subscriber systems.
15. Write about message queues.
16. Explain shared memory approach.

17. Describe distributed objects.

18. How to convert from objects to components?

19. Write about EJB.

 PEER TO PEER SERVICES AND FILE
3
SYSTEM

3.1 INTRODUCTION TO PEER TO PEER SYSTEMS

Peer-to-peer (P2P) is a decentralized communications model in which each party

has the same capabilities and either party can initiate a communication session.

Unlike the client/server model, in which the client makes a service request and the
server fulfills the request, the P2P network model allows each node to function as both
client and server.

Server Based Communication Peer to Peer Communication

Fig 3.1 Communication in Peer to peer and Client/ server model

P2P systems can be used to provide anonymized routing of network traffic, massive
parallel computing environments, distributed storage and other functions. Most P2P programs
are focused on media sharing and P2P is therefore often associated with software
piracy and copyright
violation.
 Distributed 3.2
System
Features of Peer to Peer Systems
 Large scale sharing of data and resources
 No need for centralized management
 Their design of P2P system must be in such a manner that each user contributes
resources to the entire system.
 All the nodes in apeer-to-peer system have the same functional capabilities and
responsibilities.
 The operation of P2P system does not depend on the centralized management.
 The choice of an algorithm for theplacement of data across many hosts and the
access of the data must balance the workload and ensure the availability without
much overhead.

Advantages of P2P systems over client/ server architecture

1) It is easy to install and so is the configuration of computers on the network.
2) All the resources and contents are shared by all the peers, unlike server-client
architecture where Server shares all the contents and resources.

3) P2P is more reliable as central dependency is eliminated. Failure of one peer

doesn‟t affect the functioning of other peers. In case of Client –Server network, if
server goes down whole network gets affected.
4) There is no need for full-time System Administrator. Every user is the
administrator of his machine. User can control their shared resources.
5) The over-all cost of building and maintaining this type of network is comparatively
very less.

Disadvantages of P2P systems over client/ server architecture

1) In this network, the whole system is decentralized thus it is difficult to administer.
That is one person cannot determine the whole accessibility setting of whole
network.
2) Security in this system is very less viruses, spywares,trojans, etc malwares can
easily transmitted this architecture.
3) Data recovery or backup is very difficult. Each computer should have its own back-
up system

4) Lot of movies, music and other copyrighted files are transferred using this type of
file transfer. P2P is the technology used in torrents.
 Distributed 3.3
System
P2P Middleware

 Middleware is the software that manages and supports the different components of
a distributed system.

 In essence, it sits in the middle of the system.

 Middleware is usually off Middleware is usually off-the-shelf rather than specially

written software.

 A key problem in Peer-to-Peer applications is to provide a way for clients to access

data resources efficiently.

 Peer clients need to locate and communicate with any available resource, even
though resources may be widely distributed and configuration may be dynamic,
constantly adding and removing resources and connections.

 The P2P middleware must possess the following characteristics:

 Global Scalability

 Load Balancing

 Local Optimization

 Adjusting to dynamic host availability

 Security of data

 Anonymity, deniability, and resistance to censorship

3.1.1 Routing Overlays

A routing overlay is a distributed algorithm for a middleware layer

responsible for routing requests from any client to a host that holds the object
to which the request is addressed.

 Any node can access any object by routing each request through a sequence of
nodes, exploiting knowledge at each of theme to locate the destination object.
 Global User IDs (GUID) also known as opaque identifiers are used as names, but
they do not contain the location information.
 A client wishing to invoke an operation on an object submits a request including
the object‟s GUID to the routing overlay, which routes the request to a node at
which a replica of the object
resides.
 Distributed 3.4
System

Fig 3.2: Information distribution in Routing

Overlay Differences between Overlay networks and IP routing

IP Overlay Network
The scalability of IPV4 is limited to 232 Peer to peer systems can address more
nodes and IPv6 is 2128. objects using GUID.

The load balancing is done based on Traffic patterns are independent of

topology. topology since the object locations are
randomized.
Routing tables are updated Routing tables are updated both
asynchronously. synchronously and synchronously.
Failure of one node does not degrade the Routes and object references can be
performance much. Redundancy is replicated n-fold, ensuring tolerance
introduced in IP. n-fold replication is of n failures of nodes or connections.
costly.
Each IP can map to only one node. Messages are routed to the nearest replica
of the target object.
Secure addressing is possible only Secure communication is possible
between trusted nodes. between limited trusted systems.
 Distributed 3.5
System
Distributed Computation
 Distributed computing refers to multiple computer systems working on a single
problem.
 Here a single problem is divided into many parts, and each part is solved by
different computers.
 As long as the computers are networked, they can communicate with each other to
solve the problem.
 The computers perform like a single entity.
 The ultimate goal of distributed computation is to maximize performance by
connecting users and IT resources in a cost-effective, transparent and reliable
manner.
 It also ensures fault tolerance and enables resource accessibility in the event that
one of the components fails.

3.2 NAPSTER AND ITS LEGACY APPLICATION

 The Internet was originally built as a peer-to-peer system in the late 1960s to
share computing resources within the US.
 Later the transfer of resources between systems took place by means of client
server communication.
 Later Napster was developed for peer –to-peer file sharing especially MP3 files.
 They are not fully peer-to-peer since it used central servers to maintain lists of
connected systems and the files they provided, while actual transactions were
conducted directly between machines.

3.2.1 Working of Napster

 Each user must have Napster software in order to engage in file transfers.
 The user runs the Napster program. Once executed, this program checks for an
Internet connection.
 If an Internet connection is detected, another connection between the user's
computer and one of Napster's Central Servers will be established. This connection
is made possible by the Napster file-sharing software.
 The Napster Central Server keeps a directory of all client computers connected to
it and stores information on them as described above.
 Distributed 3.6
System
 If a user wants a certain file, they place a request to the Napster Centralised Server
that it's connected to.
 The Napster Server looks up its directory to see if it has any matches for the user's
request.
 The Server then sends the user a list of all that matches (if any) it as found
including the corresponding, IP address, user name, file size, ping number, bit rate
etc.
 The user chooses the file it wishes to download from the list of matches and tries
to establish a direct connection with the computer upon which the desired file
resides.
 It tries to make this connection by sending a message to the client computer
indicating their own IP address and the file name they want to download from the
client.
 If a connection is made, the client computer where the desired file resides is now
considered the host.
 The host now transfers the file to the user.
 The host computer breaks the connection with the user computer when
downloading is complete.
peers

Napster server
Napster server
Index 1. File location
Index
request
2. List of peers
of f ering the f
ile
5. Index update
4. File deliv
ered

Fig 3.1: Napster

Limitations of Napster:
Discovery and addressing a file is difficult unless the location is distributed.
 Distributed 3.7
System
3.3 PEER TO PEER MIDDLEWARE
 Middleware is the software that manages and supports the different components
of a distributed system.
 In essence, it sits in the middle of the system.
 Middleware is usually off Middleware is usually off-the-shelf rather than
specially written software.
 A key problem in Peer-to-Peer applications is to provide a way for clients to
access data resources efficiently.
 Peer clients need to locate and communicate with any available resource, even
though resources may be widely distributed and configuration may be dynamic,
constantly adding and removing resources and connections.
 The following are the functional requirements of the middleware:

 Simplify construction of services across many hosts in wide network

 Add and remove resources at will
 Add and remove new hosts at will
 Interface to application programmers should be simple and independent
of types of distributed resources
 The following are the non functionalrequirements of the middleware:
 Global Scalability

 Load Balancing
 Optimization for local interactions between neighboring peers

 Accommodation to highly dynamic host availability

 Security of data in an environment simplify construction of
services across many hosts in wide network
 Anonymity, deniability and resistance to censorship
 Sharing and balancing across large numbers of computers pose major design
challenges to the development of middleware.
 In order to locate an object, the knowledge of object location must be distributed
throughout network through
replication.
 Distributed 3.8
System
3.4 ROUTING OVERLAY

A routing overlay is a distributed algorithm for a middleware layer responsible for

routing requests from any client to a host that holds the object to which the
request is addressed.

 Any node can access any object by routing each request through a sequence of
nodes, exploiting knowledge at each of theme to locate the destination object.
 Global User IDs (GUID) also known as opaque identifiers are used as names, but
they do not contain the location information.
 A client wishing to invoke an operation on an object submits a request including
the object‟s GUID to the routing overlay, which routes the request to a node at
which a replica of the object resides.
 They are actually sub-systems within the peer-to-peer middleware that are meant
locating nodes and objects.
 They implement a routing mechanism in the application layer.
 They ensure that any node can access any object by routing each request thru a
sequence of nodes
Features of GUID:
 They are pure names or opaque identifiers that do not reveal anything about the
locations of the objects.
 They are the building blocks for routing overlays.
 They are computed from all or part of the state of the object using a function that
deliver a value that is very likely to be unique. Uniqueness is then checked against
all other GUIDs.
 They are not human understandable.

Process of Routing Overlay

 Client submits a request including the object GUID, routing overlay routes the
request to a node at which a replica of the object resides.
 A node introduces a new object by computing its GUID and announces it to the
routing overlay.
 Note that clients can remove an object and also nodes may join and leave the
service
 Distributed 3.9
System
Types of Routing Overlays
1. DHT – Distributed Hash Tables. GUIDs are stored based on hash values.
2. DOLR – Distributed Object Location and Routing. DOLR is a layer over the DHT
that maps GUIDs and address of nodes. GUIDs host address is notified using the
Publish() operation.

3.5 PASTRY

Pastry is a generic, scalable and efficient substrate for peer-to-peer

applications. Pastry nodes form a decentralized, self-organizing and fault-
tolerant overlay network within the Internet.

 Pastry provides efficient request routing, deterministic object location, and load
balancing in an application-independent manner.
 Furthermore, Pastry provides mechanisms that support and facilitate application-
specific object replication, caching, and fault recovery.
 Each node in the Pastry network has a unique, uniform random identifier (nodeId) in
a circular 128-bit identifier space.
 When presented with a message and a numeric 128-bit key, a Pastry node efficiently
routes the message to the node with a nodeId that is numerically closest to the key,
among all currently live Pastry nodes.
 The expected number of forwarding steps in the Pastry overlay network is O(log N),
while the size of the routing table maintained in each Pastry node is only O(log N) in
size (where N is the number of live Pastry nodes in the overlay network).
 At each Pastry node along the route that a message takes, the application is notified
and may perform application-specific computations related to the message.
 Each Pastry node keeps track of its L immediateneighbors in the nodeId space called
the leaf set, and notifies applications of new node arrivals, node failures and node
recoveries within the leaf set.
 Pastry takes into account locality (proximity) in the underlying Internet.
 It seeks to minimize the distance messages travel, according to a scalar proximity
metric like the ping delay.
 Pastry is completely decentralized, scalable, and self-organizing; it automatically
adapts to the arrival, departure and failure of nodes.
 Distributed 3.10
System
Capabilities of Pastry
 Mapping application objects to Pastry nodes
 Application-specific objects are assigned unique, uniform random identifiers
(objIds) and mapped to the k, (k >= 1) nodes with nodeIds numerically closest
to the objId.
 The number k reflects the application's desired degree of replication for the
object.
 Inserting objects
 Application-specific objects can be inserted by routing a Pastry message, using
the objId as the key.
 When the message reaches a node with one of the k closest nodeIds to the
objId, that node replicates the object among the other k-1 nodes with closest
nodeIds (which are, by definition, in the same leaf set for k <= L/2).
 Accessing objects
 Application-specific objects can be looked up, contacted, or retrieved by
routing a Pastry message, using the objId as the key.
 By definition, the message is guaranteed to reach a node that maintains a
replica of the requested object unless all k nodes with nodeIds closest to the
objId have failed.
 Availability and persistence
 Applications interested in availability and persistence of application-specific
objects maintain the following invariant as nodes join, fail and recover: object
replicas are maintained on the k nodes with numerically closest nodeIds to the
objId, for k> 1.
 The fact that Pastry maintains leaf sets and notifies applications of changes in
the set's membership simplifies the task of maintaining this invariant.
 Diversity
 The assignment of nodeIds is uniform random, and cannot be corrupted by an
attacker. Thus, with high probability, nodes with adjacent nodeIds are diverse
in geographic location, ownership, jurisdiction, network attachment, etc.
 The probability that such a set of nodes is conspiring or suffers from correlated
failures is low even for modest set sizes.
 This minimizes the probability of a simultaneous failure of all k nodes that
maintain an object
replica.
 Distributed 3.11
System
 Load balancing
 Both nodeIds and objIds are randomly assigned and uniformly distributed in
the 128-bit Pastry identifier space.
 Without requiring any global coordination, this results in a good first-order
balance of storage requirements and query load among the Pastry nodes, as
well as network load in the underlying Internet.

 Object caching
 Applications can cache objects on the Pastry nodes encountered along the
paths taken by insert and lookup messages.
 Subsequent lookup requests whose paths intersect are served the cached copy.
 Pastry's network locality properties make it likely that messages routed with
the same key from nearby nodes converge early, thus lookups are likely to
intercept nearby cached objects.
 This distributed caching offloads the k nodes that hold the primary replicas of
an object, and it minimizes client delays and network traffic by dynamically
caching copies near interested clients.
 Efficient, scalable information dissemination
 Applications can perform efficient multicast using reverse path forwarding along
the tree formed by the routes from clients to the node with nodeId numerically
closest to a given objId.
 Pastry's network locality properties ensure that the resulting multicast trees are
efficient; i.e., they result in efficient data delivery and resource usage in the
underlying Internet.

3.5.1 Pastry’s Routing Algorithm

The routing algorithm involves the use of a routing table at each node to route
messages efficiently. This is divided into two stages:
 First Stage:
 In this stage a routing scheme is used, that routes messages correctly but
inefficiently without a routing table.
 Each active node stores a leaf set – a vector L (of size 2l) containing the
GUIDs and IP addresses of the nodes whose GUIDs are numerically closest on
either side of its own (l above and l below).
 Distributed 3.12
System
 Leaf sets are maintained by Pastry as nodes join and leave. Even after a node
failure, they will be corrected within a short time.
 In the Pastry system that the leaf sets reflect a recent state of the system and
that they converge on the current state in the face of failures up to some
maximum rate of failure.
 The GUID space is treated in circular fashion.
 GUID 0‟s lower neighbour is 2128–1.
 Every leaf set includes the GUIDs and IP addresses of the current
node‟s immediateneighbours.
 A Pastry system with correct leaf sets of size at least 2 can route messages to
any GUID trivially as follows: any node A that receives a messageM with
destination address D routes the message by comparing D with its own GUIDA
and with each of the GUIDs in its leaf set and forwarding M to the node
amongstthem that is numerically closest to D.

 Second Stage:
 This has a full routing algorithm, which routes a request to any node
in O(log N) messages.
 Each Pastry node maintains a tree-structured routing table with GUIDs and
IP addresses for a set of nodes spread throughout the entire range of 2128
possible GUID values, with increased density of coverage for GUIDs
numerically close to its own.
 The structure of routing table is: GUIDs are viewed as hexadecimal values and
the table classifies GUIDs based on their hexadecimal prefixes.
 The table has as many rows asthere are hexadecimal di gits in a GUID, so for
the prototype Pastry system that weare describing, there are 128/4 = 32 rows.
 Any row n contains 15 entries – one foreach possible value of the
nth hexadecimal digit, excluding the value in the local node ‟s
GUID.
 Each entry in the table points to one of the potentially many nodes
whose GUIDs have the relevant prefix.
 The routing process at any node A uses the information in its routing table R
and leaf set L to handle each request from an application and each incoming
messagefrom another node according to the algorithm.
 Distributed 3.13
System
Pastry’s Routing Algorithm
To handle a message M addressed to a node D (where R[p,i] is the element at column i,
row p of the routing table):
1. If (L-l < D <Ll) { // the destination is within the leaf set or is the current node.
2. Forward M to the element Li of the leaf set with GUID closest to D or
the current node A.
3. } else { // use the routing table to despatch M to a node with a closer GUID
4. Find p, the length of the longest common prefix ofD and A,.and i, the (p+1)th
hexadecimal digit of D.
5. If (R[p,i] • null) forward M to R[p,i] // route M to a node with a longer common prefix.
6. else { // there is no entry in the routing table.
7. Forward M to any node in L or R with a common prefix of length p but a
GUID that is numerically closer.
}
}

Locality
 The Pastry routing structure is redundant.
 Each row in a routin g table contains 16 entries.
 The entries in the ithrow give the addresses of 16 nodes with GUIDs with i–1 initial
hexadecimal digits thatmatch the current node ‟s GUID and an ith di git that takes each
of the possible hexadecimal mal values.

 A well-populated Pastry overlay will contain many more nodesthan can be contained
in an individual routing table; whenever a new routing table isbeing constructed a
choice is made for each position between several candidates based on a proximity
neighbour selection algorithm.
 A locality metric is used to compare candidates and the closest available node is
chosen.
 Since the information available is not comprehensive, this mechanism cannot
produce globally optimal routings, but simulations have shown that it results in routes
that are on average only about 30–50% longer than the optimum.
 Distributed 3.14
System
Fault tolerance
 The Pastry routing algorithm assumes that allentries in routing tables and leaf sets
refer to live, correctly functioning nodes.
 All nodessend „heartbeat‟ messages i.e., messages sent at fixed time intervals to
indicate that thesender is alive to neighbouring nodes in their leaf sets, but
information about failednodes detected in this manner may not be disseminated
sufficiently rapidly to eliminaterouting errors.
 Nor does it account for malicious nodes that may attempt to interfere with
correct routing.
 To overcome these problems, clients that depend upon reliable messagedelivery
are expected to employ an at-least-once delivery mechanism and repeat their
requests several times in the absence of a response.
 This willallow Pastry a longer time window to detect and repair node failures.
 To deal with any remaining failures or malicious nodes, a small degree of
randomness is introduced into the route selection algorithm.

Dependability
 Dependability measures include the use of acknowledgements at each hop in the
routing algorithm.
 If the sending host does not receive an acknowledgement after aspecified timeout,
it selects an alternative route and retransmits the message.
 The nodethat failed to send an acknowledgement is then noted as a suspected
failure.
 To detect failed nodes each Pastry node periodically sends aheartbeat message to
its immediate neighbour to the left (i.e., with a lower GUID) in the
leaf set.
 Each node also records the time of the last heartbeat message received from its
immediate neighbour on the right (with a higher GUID).
 If the interval since the lastheartbeat exceeds a timeout threshold, the detecting
node starts a repair procedure thatinvolves contacting the remaining nodes in the
leaf set with a notification about thefailed node and a request for suggested
replacements.
 Even in the case of multiplesimultaneous failures, this procedure terminates with
all nodes on the left side of thefailed node having leaf sets that contain the l live
nodes with the closest
GUIDs.
 Distributed 3.15
System
 Suspectedfailed nodes in routing tables are probed in a similar manner to that used
for the leaf setand if they fail to respond, their routing table entries are replaced
with a suitablealternative obtained from a nearby node.
 A simple gossip protocol is used to periodically exchange routing table
information betweennodes in order to repair failed entries and prevent slow
deterioration of the localityproperties.
 The gossip protocol is run about every 20 minutes.

3.6 TAPESTRY
 Tapestry is a decentralized distributed system.
 It is an overlay network that implements simple key-based routing.
 Each node serves as both an object store and a router that applications can contact
to obtain objects.
 In a Tapestry network, objects are “published” at nodes, and once an object has
been successfully published, it is possible for any other node in the network to find
the location at which that objects is published.
 Identifiers are either NodeIds, which refer to computers that perform routing
operations, or GUIDs, which refer to the objects.
 For any resource with GUID G there is a unique root node with GUID RG that is
numerically closest to G.
 Hosts H holding replicas of G periodically invoke publish(G) to ensure that newly
arrived hosts become aware of the existence of G.

4377

43FE 437A

4228
4361
4378 4664
Phil‟s
Books 4B4F 4A6D

E791 4378
57EC AA93
Phil‟s
Books

Fig 3.4: Tapestry Routing

 Distributed 3.16
System
Unstructured peer-to-peer Networks
 In structured approaches, there is an overall global policy governing the topology
of the network, the placement of objects in the network and the routing or search
functions used to locate objects in the network.
 There is a specific data structure underpinning the associated overlay and a set of
algorithms operating over that data structure.
 These algorithms are efficient and time bounded.
 In unstructured approaches, there is no overall control over the topology or the
placement of objects within the network.
 The overlay iscreated in an ad hoc manner, with each node that joins the network
following simple, local rules to establish connectivity.
 In particular, a joining node will establishcontact with a set of neighbours
knowing that the neighbours will also be connected tofurtherneighbours and so on,
forming a network that is fundamentally decentralized andself-organizing g and
hence resilient to node failure.
 To locate a given object, it is thennecessary to carry out a search of the resultant
network topology.
 This approachcannot offer guarantees of being able to find the object and
performance will beunpredictable.
 There is a real risk of generating excessive message traffic tolocate objects.

Differences between structured and unstructured networks

Structured Network Unstructured Network

This guarantees to locate objects and can This is self-organizing and resilient to
offer time and complexity bounds on this node failure.
operation. So it has relatively low message
overhead.
This needs complexoverlay structures, which Probabilistic and hence cannot offer
can bedifficult and costly to achieve, absolute guarantees on locating objects;
especially in highly dynamic prone to excessive messaging overhead
environments. which can affect scalability

Gnutella
 The Gnutella network is a fully decentralized, peer-to-peer application layer
network that facilitates file sharing; and is built around an open protocol developed
to enable host discovery, distributed search, and file transfer.
 Distributed 3.17
System
 It consists of the collection of Internet connected hosts on which Gnutella protocol
enabled applications are running.
 The Gnutella protocol makes possible all host-to-host communication through the
use of messages.

Fig : 3.5: Message Format

 A message consists of the following five fields:
 The GUID field provides a unique identifier for a message on the network; the
Type field indicates which type of message is being communicated.
 TTL field enumerates the maximum number of hops that this message is
allowed to traverse.
 Hops field provides a count of the hops already traversed.
 Payload Size field provides a byte count of all data expected to follow the
message.
 A host communicates with its peers by receiving, processing, and forwarding
messages, but to do so it must follow a set of rules which help to ensure
reasonable message lifetimes.
 The rules are as follows:
1. Prior to forwarding a message, a host will decrement its TTL field and
increment its Hops field. If the TTL field is found to be zero following this
action, the message is not forwarded.
2. If a host encounters a message with the same GUID and Type fields as a
message already forwarded, the new message is treated as a duplicate and is
not forwarded a second time.

Types of Message
There are only five types of messages
 Ping
 Pong
 Query
 Distributed 3.18
System
 Query-Hit
 Push
Ping and Pong messages facilitate host discovery, Query and Query-Hit messages make
possible searching the network, and Push messages ease file transfer from firewalled
hosts.Because there is no central index providing a list of hosts connected to the network, a
disconnected host must have offline knowledge of a connected host in order to connect to the
network.
 Once connected to the network, a host is always involved in discovering hosts,
establishing connections, and propagating messages that it receives from its peers,
but it may also initiate any of the following six voluntary activities: searching for
content, responding to searches, retrieving content, and distributing content. A host
will typically engage in these activities simultaneously.

 A host will search for other hosts and establish connections so as to satisfy a
maximum requirement for active connections as specified by the user, or to replace
active connections dropped by it or its peer.
 Consequently, a host tends to always maintain the maximum requirement for
active connections as specified by the user, or the one connection that it needs to
remain connected to the network.
 To engage in host discovery, a host must issue a Ping message to the host of which
it has offline knowledge.
 That host will then forward the ping message across its open connections, and
optionally respond to it with a Pong message.
 Each host that subsequently receives the Ping message will act in a similar manner
until the TTL of the message has been exhausted.
 A Pong message may only be routed along the reverse of the path that carried the
Ping message to which it is responding.
 After having discovered a number of other hosts, the host that issued the initial
ping message may begin to open further connections to the network.
 Doing so allows the host to issue ping messages to a wider range of hosts and
therefore discover a wider range of other hosts, as well as to begin querying the
network.
 In searching for content, a host propagates search queries across its active
connections.
 Those queries are then processed and forwarded by its peers.
 Distributed 3.19
System
 When processing a query, a host will typically apply the query to its local database
of content, and respond with a set of URLs pointing to the matching files.
 The propagation of Query and Query-Hit messages is identical to that of Ping and
Pong messages.
 A host issues a Query message to the hosts to which it is connected.
 The hosts receiving that Query message will then forward it across their open
connections, and optionally respond with a Query-Hit message.
 Each host that subsequently receives the Query message will act in a similar
manner until the TTL of the message has been exhausted.
 Query-Hit messages may only be routed along the reverse of the path that carried
the Query message to which it is a response.
 The sharing of content on the Gnutella network is accomplished through the use of
the HTTP protocol.
 Due to the decentralized nature of the architecture, each host participating in the
Gnutella network plays a key role in the organization and operation of the
network.
 Both the sharing of host information, as well as the propagation of search requests
and responses are the responsibility of all hosts on the network rather than a central
index.

Fig 3.6: Elements in Gnutella Network

 The Gnutella network architecture avoids creating the single-point-of-failure that,
theoretically, threatens other
networks.
 Distributed 3.20
System
 The Gnutella network exhibits a great deal of fault-tolerance, continuing to
function even as subsets of hosts become unavailable.

 Each host in the Gnutella network is capable of acting as both a server and client,
allowing a user to distribute and retrieve content simultaneously.

 This spreads the provision of content across many hosts, and helps to eliminate the
bottlenecks that typically occur when placing enormous loads on a single host.

 In this network, the amount and variety of content available to a user scales with
the number of users participating in the network.

3.7 DISTRIBUTED FILE SYSTEMS

 The purpose of a distributed file system (DFS) is to allow users of physically
distributed computers to share data and storage resources by using a common file
system.

 A typical configuration for a DFS is a collection of workstations and mainframes

connected by a local area network (LAN).

 A DFS is implemented as part of the operating system of each of the connected

computers.

Distributed file systems support the sharing of information in the form of files
and hardware resources.

 The following are the modules of a file systems:

 Directory module: Relates file names to the disk.
 File module: Relates file ID to the files
 Access Control module: Manages access rights
 File access module: Controls read or write operation on the files
 Block module: Allocates disk blocks and helps in accessing them
 Device module: Manages disk I/O and buffering.
 Distributed 3.21
System

Fig 3.6: File Attribute Record Structure

File Attribute Record Structure
 File length specifies the size of the file.
 Creation Timestamp is the time at which the file was created.
 Read Timestamp is the time at which the last read operation was made.
 Write Timestamp is the time at which the last write operation was made.
 Reference Count is the number or links to the file.
 Owner is the creator of the file who has full access rights.
 File type describes the content of the file.
 Access Control List contains the list of users and the access rights given to them.
3.7.1 Unix File Systems
The time sharing Unix file system is a model for the distributed file systems. The
following some of the unix commands to access files:
 filedes= open(name, mode): Opens an already existing file.
 filedes=creat(name, mode): Creates a new file
 status= close(filedes): Closes a file.
 count= read(filedes, buffer, n): Transfers n bytes from the filedes to buffer and
returns the number of bytes transferred.
 count= write(filedes, buffer, n): Transfers n bytes from the filedes to buffer and
returns the number of bytes transferred. This advances the write file pointer.
 Distributed 3.22
System
 pos= lseek(filedes, offset, whence): Moves the read-write pointer to offset
depending on whence.
 status=unlink(name): renames the specified file.
 status=unlink(name1, name2): Adds a new name (name2) for a file (name1).
 status= stat(name, buffer): Gets the file attributes for file into the buffer.
3.7.2 Characteristics of a File system
 File systems are responsible for the organization, storage, retrieval, naming,
sharing and protection of files.
 Data and attributes are the primary characteristics of a file system.
 The data consist of a sequence of data items .
 The attributes are held as a single record containing information such as the
length of the file, timestamps, file type, owner‟s identity and access control lists.
 A directory is a special file that provides a mapping from text names to internal
file identifiers.
 The term metadata is used to refer to all of the extra information stored by
a file system that is needed for the management of files.
 It includes file attributes , directories and all the other persistent information used
by the file system.
 The file system also maintains the access control lists.
3.7.3 Requirements of distributed file systems
Transparency is operating in such a way as to not be perceived by users. The distributed
file system demands the following transparency requirements:
 Login Transparency: User can log in at any host with uniform login procedure
and perceive a uniform view of the file system.
 Access Transparency: Client process on a host has uniform mechanism to access
all files in system regardless of files are on local/remote host.
 Location Transparency: The names of the files do not reveal their physical
location.
 Concurrency Transparency: An update to a file should not have effect on the
correct execution of other process that is concurrently sharing a file.
 Replication Transparency: Files may be replicated to provide redundancy for
availability and also to permit concurrent access for efficiency.
 Distributed 3.23
System
Apart from the above transparency properties, the file systems also need the following:
 Fault Tolerance: It is a design that enables a system to continue operation,
possibly at a reduced level, rather than failing completely, when some part
of the system fails.
 Network Transparency: Same access operation as if they are local files.
 Location Independence: The file name should not be changed when the physical
location of the file changes.
 User Mobility: User should be able to access the file from any where.
 File Mobility: Moves files from one place to the other in a running system.

Examples:
 The GOOGLE File System: A scalable distributed file system for large
distributed data-intensive applications. It provides fault tolerance while running on
inexpensive commodity hardware, and it delivers high aggregate performance to a
large number of clients.
 The CODA distributed file system: Developed at CMU, incorporates many
distinguished features which are not present in any other distributed file system.

3.8 FILE SERVICE ARCHITECTURE

File service architecture offers a clear separation of the main concerns in providing
access to files is obtained by structuring the file service as three components:
 A flat file service
 A directory service

 A client module.
Client Computer Server Computer

Directory service
Application Application
Program Program

Flat file service

Client module

Fig 3.9: File Service Architecture

 Distributed 3.24
System
Flat file service:

 This is concerned with the implementation of operations on the contents of file.

 Unique File Identifiers (UFIDs) are used to refer to files in all requests for flat file
service operations.

 UFIDs are long sequences of bits chosen so that each file has a unique among all
of the files in a distributed system.

Directory service:

 This provides mapping between text names for the files and their UFIDs.

 Clients may obtain the UFID of a file by quoting its text name to directory service.

 Directory service supports functions needed generate directories, to add new files
to directories.

Client module:

 It runs on each computer and provides integrated service (flat file and directory) as
a single API to application programs.

 It holds information about the network locations of flat-file and directory server
processes; and achieve better performance through implementation of a cache of
recently used file blocks at the client.

The following are the operations in file service:

 Read(FileId, i, n)Data :Reads a sequence of up to n items from a file starting at

item i and returns it in Data.

 Write(FileId, i, Data):Write a sequence of Data to a file, starting at item i,

extending the file if necessary.

 Create() FileId: Creates a new file of length0 and delivers a UFID for it.

 Delete(FileId): Removes the file from the file store.

 GetAttributes(FileId) Attr:Returns the file attributes for the file.

 SetAttributes(FileId, Attr):Sets the file attributes .

 Access control in files:

In distributed implementations, access rights checks have to be performed at the server

because the server RPC interface is an otherwise unprotected point of access to files.
 Distributed 3.25
System
 Hierarchic file system
A hierarchic file system such as the one that UNIX provides consists of a number of
directories arranged in a tree structure.
 File Group
 A file group is a collection of files that can be located on any server or moved between
servers while maintaining the same names.
 A similar construct is used in a UNIX file system.
 It helps with distributing the load of file serving between several servers.

 File groups have identifiers which are unique throughout the system.
 To construct a globally unique ID we use some unique attribute of the machine on
which it is created, e.g. IP number, even though the file group may move
subsequently.

Fig 3.9: File group ID

3.9 ANDREW FILE SYSTEM (AFS)
– This was developed by Carnegie Mellon University as part of Andrew distributed
computing environments (in 1986).

– The public domain implementation is available on Linux (LinuxAFS)

– It was adopted as a basis for the DCE/DFS file system in the Open Software
Foundation (OSF, www.opengroup.org) DEC (Distributed Computing Environment).
– Like NFS, AFS provides transparent access to remote shared files for UNIX
programs running on workstations.
– AFS is implemented as two software components that exist at UNIX processes called
Vice and Venus.
– The files available to user processes running on workstations are either local or
shared.
– Local files are handled as normal UNIX files.

– They are stored on the workstation‟s disk and are available only to local user
processes.
 Distributed 3.26
Workstations System
Venus
Servers
User
program
UNIX Kernel Vice

UNIX Kernel
Venus
User
program Network
UNIX Kernel

Vice
Venus UNIX Kernel
User
program
UNIX Kernel

Fig 3.10: Process distribution in AFS

– Shared files are stored on servers, and copies of them are cached on the local disks of
workstations.
– The UNIX kernel in each workstation and server is a modified version of BSD
UNIX.
– The modifications are designed to intercept open, close and some other file system
calls when they refer to files in the shared name space and pass them to the Venus
process in the client computer.

Workstation

Non-local file
User operations Venus
program UNIX file
system calls

UNIX Kernel
UNIX file system

Local
disk

Fig 3.11: System call interception

 Distributed 3.27
System
Components of Vice service interface
 Fetch(fid)attr, data: Returns the attributes (status) and, optionally, the contents of
the file identified by fid and records a callback promise on it.
 Store(fid, attr, data): Updates the attributes and (optionally) the contents of a
specified file.
 Create() fid: Creates a new file and records a callback promise on it.

 Remove(fid) Deletes the specified file.

 SetLock(fid, mode): Sets a lock on the specified file or directory. The mode of the
lock may be shared or exclusive. Locks that are not removed expire after 30
minutes.
 ReleaseLock(fid):Unlocks the specified file or directory.
 RemoveCallback(fid): Informs the server that a Venus process has flushed a file
from
its cache.
 BreakCallback(fid:) Call made by a Vice server to a Venus process; cancels the
call back promise on the relevant file.

3.10 FILE ACCESSING MODELS

File access models are methods used for accessing remote files and the unit of data
access. A distributed file system may use one of the following models to service a client‟s file
access request when the accessed file is remote:

1. Remote service model

 Processing of a client‟s request is performed at the server‟s node.

 Thus, the client‟s request for file access is delivered across the network as a
message to the server, the server machine performs the access request, and the
result is sent to the client.
 This need to minimize the number of messages sent and the overhead per
message.

2. Data-caching model
 This model attempts to reduce the network traffic of the previous model by
caching the data obtained from the server node.
 This takes advantage of the locality feature of the found in file accesses.
 Distributed 3.28
System
 A replacement policy such as LRU is used to keep the cache size bounded.
 This model reduces network traffic it has to deal with the cache coherency
problem during writes, because the local cached copy of the data needs to be
updated, the original file at the server node needs to be updated and copies in
any other caches need to be updated.
 The data-caching model offers the possibility of increased performance and
greater system scalability because it reduces network traffic, contention for the
network, and contention for the file servers. Hence almost all distributed file
systems implement some form of caching.
 In file systems that use the data-caching model, an important design issue is to
decide the unit of data transfer.
 This refers to the fraction of a file that is transferred to and form clients as a
result of single read or write operation.
3. File-level transfer model
 In this model when file data is to be transferred, the entire file is moved.
 File needs to be transferred only once in response to client request and hence is
more efficient than transferring page by page which requires more network
protocol overhead.
 This reduces server load and network traffic since it accesses the server only
once. This has better scalability.
 Once the entire file is cached at the client site, it is immune to server and
network failures.
 This model requires sufficient storage space on the client machine.
 This approach fails for very large files, especially when the client runs on a
diskless workstation.
 If only a small fraction of a file is needed, moving the entire file is wasteful.

4. Block-level transfer model

 File transfer takes place in file blocks.
 A file block is a contiguous portion of a file and is of fixed length (can also be
a equal to a virtual memory page size).
 This does not require client nodes to have large storage space.
 It eliminates the need to copy an entire file when only a small portion of the
data is
needed.
 Distributed 3.29
System
 When an entire file is to be accessed, multiple server requests are needed,
resulting in more network traffic and more network protocol overhead. NFS
uses block-level transfer model.

5. Byte-level transfer model

 Unit of transfer is a byte.
 Model provides maximum flexibility because it allows storage and retrieval of
an arbitrary amount of a file, specified by an offset within a file and length.
 The drawback is that cache management is harder due to the variable-length
data for different access requests.
6. Record-level transfer model
 This model is used with structured files and the unit of transfer is the record.

3.11 FILE-SHARING SEMANTICS

 Multiple users may access a shared file simultaneously.
 An important design issue for any file system is to define when modifications of
file data made by a user are observable by other users.
 The UNIX semantics is implemented in file systems for single CPU systems
because it is the most desirable semantics and because it is easy to serialize all
read/write requests.
 Implementing UNIX semantics in a distributed file system is not easy.
 One may think that this can be achieved in a distributed system by disallowing
files to be cached at client nodes and allowing a shared file to be managed by only
one file server that processes all read and write requests for the file strictly in the
order in which it receives them.
 However, even with this approach, there is a possibility that, due to network
delays, client requests from different nodes may arrive and get processed at the
server node in an order different from the actual order in which the requests were
made.
 Also, having all file access requests processed by a single server and disallowing
caching on client nodes is not desirable in practice due to poor performance, poor
scalability, and poor reliability of the distributed file system.
 Hence distributed file systems implement a more relaxed semantics of file sharing.
 Applications that need to guarantee UNIX semantics should provide mechanisms
(e.g. mutex lock etc) themselves and not rely on the underlying semantics of
sharing provided by the file
system.
 Distributed 3.30
System
3.12 NAMING IN DISTRIBUTED SYSTEMS
 A Name is a string of bits used to refer to an entity.
 We operate on an entity through its Access Point. The Address is the name of the
access point.

 The naming faciliy of a distributed operating system enables users and programs to
assign character-string names to objects and subsequently use these names to refer
to those objects.
 The locating faciliy, which is an integral part of the naming facility, maps an
object's name to the object's location in a distributed system.

 The naming and locating facilities jointly form a naming system that provides the
users with an abstraction of an object that hides the details of how and where an
object is actually located in the network.

 It provides a further level of abstraction when dealing with object replicas.

 Given an object name, it returns a set of the locations of the object's replicas.

 The naming system plays a very important role in achieving the goal of

 location transparency,

 facilitating transparent migration

 replication of objects, v object sharing.

Example:

 Telephone as Access Point to a person.

 The Telephone Number then becomes the address of the person.

 Person can have several telephone numbers.

 Entity can have several addresses

3.12.1 Identifiers:

– Identifiers are special names that uniquely identify an entity.

– An identifier refers to at most one entity.

– Each entity is referred to at most one identifier.

– An identifier always refers to the same entity (never reused).

 Distributed 3.31
System
3.12.2 Namespaces:

Fig 3.12: Namespaces

– Names are organized into Name Space.
– Entities in a structured name space are named by a path name.
– Leaf nodes represent named entities (e.g., files) and have only incoming edges .
– Directory nodes have named outgoing edges and define the path used to find a leaf
node
Attribute based naming
• This allows a user to search for an entity whose name is not known.
• Entities are associated with various attributes, which can have specific values.
• By specifying a collection of <attribute, value> pairs, a user can identify one (or
more) entities
• Attribute based naming systems are also referred to as directory services, as
opposed to naming systems.
3.12.3 Naming resolution
It is the process of looking up a name. The path name of the file is given
as N:<label-1, label-2,…, label-n>

Name resolution is the process of mapping an object'sname to the object's

properties, such as its location.
 Distributed 3.32
System
 Since an object's properties are stored and maintained by the authoritative name
servers of that object, name resolution is basically the process of mapping an
object's name to the authoritative name servers of that object.
 Once an authoritative name server of the object has been located, operations can be
invoked to read or update the object's properties.
 Each name agent in a distributed system knows about at least one name server
apriori.
 To get a name resolved, a client first contacts its name agent, which in turn
contacts a known name server, which may in turn contact other name servers.
Absolute and Relative Names
– A current working context is also known by the shorter names current context or
working context.
– According to this concept, a user is always associated with a context that is his or
her current context.
– A user can change his or her current context whenever he or she desires.

Fig 3.12: Tree structured name space

– An absolute name begins at the root context of the name space tree and follows a
path down to the specified object, giving the context names on the path.
 Distributed 3.33
System
– On the other hand, a relative name defines a path from the current context to the
specified object. It is called a relative name because it is "relative to” (start from)
the user's current context.
In this method, a user may specify an object in any of the following ways:
1. Using the full (absolute) name
2. Using a relative name

3. Changing the current context first and then using a relative name

Implementing Name Spaces

Three layers used to implement such distributed name spaces
 Global layer: root node and its children
 Administrational layer: directory nodes within a single organization

 Managerial layer

Item Global Administrational Managerial

Geographical scale of network Worldwide Organization Department
Total number of nodes Few Many Vast numbers
Responsiveness of lookups Seconds Milliseconds Immediate
Updata propagation Lazy Immediate Immediate
Number of replicas Many None or few None
Is client-side caching applied? Yes Yes Sometimes

Name Space Distribution

Iterative name
resolution
– Recursive name resolution puts a higher performance demand on each name
server.
– Too high for global layer name servers
– In this method, the name agent forwards the name resolution request to the name
server that stores the first context needed to start the resolution of the given name.
– After this, the name servers that store the contexts of the given pathname are
recursively activated one after another until the authority attribute of the named
 Distributed 3.34
System
object is extracted from the context corresponding to the last component name of
the pathname.
– The last name server returns the authority attribute to its previous name server,
which then returns it to its own previous name server, and so on.
– Finally, the fast name server that received the request from the name agent returns
the authority attribute to the name agent

Advantages of recursive name resolution:

– Caching is more effective

– Communication costs may be reduced

Fig 3.12: Principle of iterative name resolution

Fig 3.13: Principle of iterative name resolution

 Distributed 3.35
System
Recursive Name resolution:

– In this method, name servers do not call each other directly.

– Rather, the name agent retains control over the resolution process and one by one
calls each of the servers involved in the resolution process.

– To continue the name resolution, the name agent sends a name resolution request
along with the unresolved portion of the name to the next name server.

– The process continues until the name agent receives the authority attribute of the
named object

3.12.4: Name Caches

Caching is an important technique for building responsive, scalable distributed

systems. A cache can be maintained either by the client or the server or by both.

Types of Name caches

 Client name caches

Caching at the client is an effective way of pushing the processing workload from
the server out to client devices,if a client has the capacity.

 Server based cache

If result data is likely to be reused by multiple clients or if the client devices do not
have the capacity then caching at the server is more effective.

 Multi-level caching

Caches can be maintained at multiple levels. For example, caches can be

maintained at all clients and all servers. Use of a cache at one level reduces the
number of requests handled at the levels below.

3.12.5: Lightweight Directory Access Protocol (LDAP)

LDAP is a standard technology for building computer network directories. A network

directory is a specialized database that stores information about devices, applications, people
and other aspects of a computer
network.
 Distributed 3.36
3.36 Peer to Peer Services and File System
System
Fig 3.14: LDAP Tree structure
Working of LDAP:
– Client starts an LDAP session by connecting to an LDAP server.
– The default port on TCP is 389.

– Client sends operation requests to the server and the server sends responses in turn.
– With some exceptions the client need not wait for a response before sending the
next request. Server may send the responses in any order.

– The client may request the following operations:

 Start TLS : Optionally protect the connection with Transport Layer Security
(TLS), to have a more secure connection
 Bind - authenticate and specify LDAP protocol version
 Search - search for and/or retrieve directory entries
 Compare - test if a named entry contains a given attribute value
 Add a new entry
 Delete an entry
 Modify an entry
 Modify Distinguished Name (DN) - move or rename an entry
 Abandon - abort a previous request
 Distributed 3.37
System
 Extended Operation - generic operation used to define other operations
 Unbind - close the connection (not the inverse of Bind)
– In addition the server may send "Unsolicited Notifications" that are not responses
to any request, e.g. before it times out a connection.

Directory Structure
 Directory is a tree of directory entries.
 Each entry consists of a set of attributes.
 An attribute has: a name , an attribute type or attribute description and one or
more values
 Attributes are defined in a schema.
 Each entry has a unique identifier called Distinguished Name (DN).
 The DN consists of its Relative Distinguished Name (RDN) constructed from
some attribute(s) in the entry.
 It is followed by the parent entry's DN.
 Think of the DN as a full filename and the RDN as a relative filename in a folder.
 DN may change over the lifetime of the entry.
 To reliably and unambiguously identify entries, a UUID might be provided in the
set of the entry's operational attributes.

File Models in Distributed Operating System

Several machines are utilized in Distributed File Systems (DFS) to supply the file system's facility.
Various file systems frequently use various conceptual models. Models based on structure and
mobility is frequently used for file modeling. In this article, you will learn about the file models in
the distributed operating system.

Types of Files models in the distributed operating systems

 Distributed 3.38
System

There are mainly two types of file models in the distributed operating system.

1. Structure Criteria
2. Modifiability Criteria

Structure Criteria

There are two types of file models in structure criteria. These are as follows:

1. Structured Files
2. Unstructured Files

Structured Files

The Structured file model is presently a rarely used file model. In the structured file model, a file is
seen as a collection of records by the file system. Files come in various shapes and sizes and with a
variety of features. It is also possible that records from various files in the same file system have
varying sizes. Despite belonging to the same file system, files have various attributes. A record is the
smallest unit of data from which data may be accessed. The read/write actions are executed on a set
of records. Different "File Attributes" are provided in a hierarchical file system to characterize the
file. Each attribute consists of two parts: a name and a value. The file system used determines the
file attributes. It provides information on files, file sizes, file owners, the date of last modification,
the date of file creation, access permission, and the date of last access. Because of the varied access
rights, the Directory Service function is utilized to manage file attributes.

The structured files are also divided into two types:

1. Files with Non-Indexed records

2. Files with Indexed records
 Distributed 3.39
Files with Non-Indexed records System

Records in non-indexed files are retrieved based on their placement inside the file. For instance, the
second record from the starting and the second from the end of the record.

Files with Indexed records

Each record contains a single or many key fields in a file containing indexed records, each of which
may be accessed by specifying its value. A file is stored as a B-tree or similar data structure or hash
table to find records quickly.

Unstructured Files

It is the most important and widely used file model. A file is a group of unstructured data sequences
in the unstructured model. Any substructure does not support it. The data and structure of each file
available in the file system is an uninterrupted sequence of bytes such as UNIX or DOS. Most latest
OS prefer the unstructured file model instead of the structured file model due to sharing of files by
multiple apps. It has no structure; therefore, it can be interpreted in various ways by different
applications.

Modifiability Criteria

There are two files model in the Modifiability Criteria. These are as follows:

1. Mutable Files
2. Immutable Files

Mutable Files

The existing operating system employs the mutable file model. A file is described as a single series
of records because the same file is updated repeatedly once new material is added. After a file is
updated, the existing contents are changed by the new contents.

Immutable Files

The Immutable file model is used by Cedar File System (CFS). The file may not be modified once
created in the immutable file model. Only after the file has been created can it be deleted. Several
versions of the same file are created to implement file updates. When a file is changed, a new file
version is created. There is consistent sharing because only immutable files are shared in this file
paradigm. Distributed systems allow caching and replication strategies, overcoming the limitation of
many copies and maintaining consistency. The disadvantages of employing the immutable file
model include increased space use and disc allocation activity. CFS uses the "Keep" parameter to
keep track of the file's current version number. When the parameter value is 1, it results in the
production of a new file version. The previous version is erased, and the disk space is reused for a
new one. When the parameter value is greater than 1, it indicates the existence of several versions of
a file. If the version number is not specified, CFS utilizes the lowest version number for actions such
as "delete" and the highest version number for other activities such as "open".
 Distributed 3.40
System
REVIEW QUESTIONS
PART - A

1. Define peer-to-peer communications.

Peer-to-peer (P2P) is a decentralized communications model in which each party has
the same capabilities and either party can initiate a communication session.
2. Give the features of Peer to Peer Systems
 Large scale sharing of data and resources
 No need for centralized management
 Their design of P2P system must be in such a manner that each user contributes
resources to the entire system.
 All the nodes in apeer-to-peer system have the same functional capabilities and
responsibilities.
 The operation of P2P system does not depend on the centralized management.
 The choice of an algorithm for theplacement of data across many hosts and the
access of the data must balance the workload and ensure the availability without
much overhead.
3. List the advantages of P2P systems over client/ server architecture
1. It is easy to install and so is the configuration of computers on the network.
2. All the resources and contents are shared by all the peers, unlike server-client
architecture where Server shares all the contents and resources.
3. P2P is more reliable as central dependency is eliminated. Failure of one peer
doesn‟t affect the functioning of other peers. In case of Client –Server network,
if server goes down whole network gets affected.
4. There is no need for full-time System Administrator. Every user is the
administrator of his machine. User can control their shared resources.
5. The over-all cost of building and maintaining this type of network is
comparatively very less.
4. List the disadvantages of P2P systems over client/ server architecture
1. In this network, the whole system is decentralized thus it is difficult to administer.
That is one person cannot determine the whole accessibility setting of whole
network.
2. Security in this system is very less viruses, spywares,trojans, etc malwares can
easily transmitted this
architecture.
 Distributed 3.41
System
3. Data recovery or backup is very difficult. Each computer should have its own
back-up system
4. Lot of movies, music and other copyrighted files are transferred using this type of
file transfer. P2P is the technology used in torrents.
5. Give the characteristics of peer-to-peer middleware.
 The P2P middleware must possess the following characteristics:
 Global Scalability
 Load Balancing
 Local Optimization
 Adjusting to dynamic host availability
 Security of data
 Anonymity, deniability, and resistance to censorship
6. Define routing overlay.
A routing overlay is a distributed algorithm for a middleware layer responsible for
routing requests from any client to a host that holds the object to which the request is
addressed.
7. Give the differences between Overlay networks and IP routing

IP Overlay Network
The scalability of IPV4 is limited to 232 Peer to peer systems can address more
nodes and IPv6 is 2128. objects using GUID.

The load balancing is done based on Traffic patterns are independent of

topology. topology since the object locations are
randomized.
Routing tables are updated Routing tables are updated both
asynchronously. synchronously and synchronously.
Failure of one node does not degrade the Routes and object references can be
performance much. Redundancy is replicated n-fold, ensuring tolerance
introduced in IP. n-fold replication is of n failures of nodes or connections.
costly.
Each IP can map to only one node. Messages are routed to the nearest replica
of the target object.
Secure addressing is possible only Secure communication is possible
between trusted nodes. between limited trusted systems.
 Distributed 3.42
System
8. What is Napster?
Napster was developed for peer –to-peer file sharing especially MP3 files.They are not
fully peer-to-peer since it used central servers to maintain lists of connected systems
and the files they provided, while actual transactions were conducted directly between
machines.
9. Give the features of GUID.
 They are pure names or opaque identifiers that do not reveal anything about the
locations of the objects.
 They are the building blocks for routing overlays.
 They are computed from all or part of the state of the object using a function that
deliver a value that is very likely to be unique. Uniqueness is then checked against
all other GUIDs.
 They are not human understandable.
10. Give the types of routing overlays.
DHT – Distributed Hash Tables. GUIDs are stored based on hash values.
DOLR – Distributed Object Location and Routing. DOLR is a layer over the DHT that
maps GUIDs and address of nodes. GUIDs host address is notified using the Publish()
operation.
11. Define pastry.
Pastry is a generic, scalable and efficient substrate for peer-to-peer applications. Pastry
nodes form a decentralized, self-organizing and fault-tolerant overlay network within
the Internet.
12. Give the Capabilities of Pastry
 Mapping application objects to Pastry nodes
 Inserting objects
 Accessing objects
 Availability and persistence
 Diversity
 Load balancing
 Object caching
 Efficient, scalable information dissemination
 Distributed 3.43
System
13. What is tapestry?
Tapestry is a decentralized distributed system.It is an overlay network that implements
simple key-based routing. Each node serves as both an object store and a router that
applications can contact to obtain objects.

14. Give the differences between structured and unstructured networks

Structured Network Unstructured Network

This guarantees to locate objects and This is self-organizing and resilient
can offer time and complexity bounds to node failure.
on this operation. So it has relatively
low message overhead.
This needs complexoverlay structures, Probabilistic and hence cannot offer
which can bedifficult and costly to absolute guarantees on locating
achieve, especially in highly dynamic objects; prone to excessive
environments. messaging overhead which can
affect scalability

15. What is Gnutella?

The Gnutella network is a fully decentralized, peer-to-peer application layer network
that facilitates file sharing; and is built around an open protocol developed to enable
host discovery, distributed search, and file transfer.

16. What is DFS?

The purpose of a distributed file system (DFS) is to allow users of physically
distributed computers to share data and storage resources by using a common file
system. A typical configuration for a DFS is a collection of workstations and
mainframes connected by a local area network (LAN). Distributed file systems support
the sharing of information in the form of files and hardware resources.

17. What are the requirements of distributed file systems?

Transparency is operating in such a way as to not be perceived by users. The
distributed file system demands the following transparency requirements:
 Login Transparency: User can log in at any host with uniform login procedure and
perceive a uniform view of the file system.
 Access Transparency: Client process on a host has uniform mechanism to access
all files in system regardless of files are on local/remote host.
 Location Transparency: The names of the files do not reveal their physical
location.
 Distributed 3.44
System
 Concurrency Transparency: An update to a file should not have effect on the
correct execution of other process that is concurrently sharing a file.
 Replication Transparency: Files may be replicated to provide redundancy for
availability and also to permit concurrent access for efficiency.
18. What are the components in file structures?
File service architecture offers a clear separation of the main concerns in providing
access to files is obtained by structuring the file service as three components:
 A flat file service
 A directory service
 A client module.
19. What is AFS?
This was developed by Carnegie Mellon University as part of Andrew distributed
computing environments (in 1986).The public domain implementation is available on
Linux (LinuxAFS). It was adopted as a basis for the DCE/DFS file system in the Open
Software Foundation (OSF, www.opengroup.org) DEC (Distributed Computing
Environment).Like NFS, AFS provides transparent access to remote shared files for
UNIX programs running on workstations
20. Define file access modes.
File access models are methods used for accessing remote files and the unit of data
access.
21. What is naming and locating facility?
The naming faciliy of a distributed operating system enables users and programs to
assign character-string names to objects and subsequently use these names to refer to
those objects. The locating faciliy, which is an integral part of the naming facility,
maps an object's name to the object's location in a distributed system.
22. Define naming resolution.
Name resolution is the process of mapping an object'sname to the object's properties,
such as its location.
23. What is absolute name?
An absolute name begins at the root context of the name space tree and follows a path
down to the specified object, giving the context names on the path.
24. What is relative name?
A relative name defines a path from the current context to the specified object. It is
called a relative name because it is "relative to” (start from) the user's current context.
 Distributed 3.45
System
25. What are the layers in name spaces?
Three layers used to implement such distributed name spaces
 Global layer: root node and its children
 Administrational layer: directory nodes within a single organization
 Managerial layer

26. What is LDAP?

LDAP is a standard technology for building computer network directories. A network
directory is a specialized database that stores information about devices, applications,
people and other aspects of a computer network.

PART – B
1. Explain about peer to peer communication.
2. Explain the working of routing overlays.
3. Describe Napster.
4. Write in detail about peer to peer middleware.
5. Explain about pastry.
6. Describe tapestry.

7. Write in detail about distributed file system.

8. Describe file service architecture.

9. Write in detail about Andrew file system.

10. Describe the file accessing models.
11. Elucidate the file sharing semantics.
12. Describe the naming in distributed systems.
 Distributed 3.46

4
System

SYNCHRONIZATION AND
REPLICATION
4.1 INTRODUCTION TO SYNCHRONIZATION
Synchronizing data in a distributed system is an enormous challenge in and of itself.
In single CPU systems, critical regions, mutual exclusion, and other synchronization problems
are solved using methods such as semaphores. These methods will not work in distributed
systems because they implicitly rely on the existence of shared memory.
Examples:
 Two processes interacting using a semaphore must both be able to access the
semaphore. In a centralized system, the semaphore is stored in the kernel and accessed
by the processes using system calls.
 If two events occur in a distributed system, it is difficult to determine which event
occurred first.
Communication between processes in a distributed system can have unpredictable
delays, processes can fail, messages may be lost synchronization in distributed systems is
harder than in centralized systems because the need for distributed algorithms. The following
are the properties of distributed algorithms:
 The relevant information is scattered among multiple machines.
 Processes make decisions based only on locally available information.
 A single point of failure in the system should be avoided.
 No common clock or other precise global time source exists

4.1.1 Issues in Clocks

Physical clocks in computers are realized as crystal oscillation counters at the
hardware level. Clock skew(offset) is common problem with clocks. Clock skew is defined as
the difference between the times on two clocks and Clock drift is the count time at different
rates. Clock drift rate is the difference in precision between a prefect reference clock and a
physical

clock.
 Distributed Systems 4.2

4.1.2 Synchronising clocks

There are two ways to synchronise a clock:

 External synchronization
 This method synchronize the process‟s clock with an authoritative external
reference clock S(t) by limiting skew to a delay bound D > 0 - |S(t) - Ci(t) | <
D for all t.
 For example, synchronization with a UTC (Coordinated Universal
Time)source.
 Internal synchronization
 Synchronize the local clocks within a distributed system to disagree by not
more than a delay bound D > 0, without necessarily achieving external
synchronization - |Ci(t) - Cj(t)| < D for all i, j, t n
 For a system with external synchronization bound of D, the internal
synchronization is bounded by 2D.
Checking the correctness of a clock
 If drift rate falls within a bound r > 0, then for any t and t‟ with t‟ > t the following
error bound in measuring t and t‟ holds:
n (1-r)(t‟-t) ≤ H(t‟) - H(t) ≤ (1+r)(t‟-t) n

 At this condition, no jumps in hardware clocks allowed

 The most frequently used conditions are:

 Monotonically increasing
 Drift rate bounded between synchronization points

 Clock may jump ahead at synchronization points

Working of Computer timer:

 To implement a clock in a computer a counter register and a holding register are used.
 The counter is decremented by a quartz crystals oscillator.
 When it reaches zero, an interrupted is generated and the counter is reloaded from the
holding register.
 Distributed Systems 4.3

Clock skew problem

To avoid the clock skew problem, two types of clocks are used:
 Logical clocks : to provide consistent event ordering
 Physical clocks : clocks whose values must not deviate from the real time by more
than a certain amount.
Software based solutions for synchronising clocks
The following techniques are used to synchronize clocks:
 time stamps of real-time clocks
 message passing
 round-trip time (local measurement)
Based on the above mentioned techniques, the following algorithmsprovides clock
synchronization:
 Cristian‟s algorithm
 Berkeley algorithm

 Network time protocol (Internet)

4.1.3 Cristian’s algorithm

Cristian suggested the use of a time server, connected to a device that receives signals
from a source of UTC, to synchronize computers externally. Round trip times between
processes are often reasonably short in practice, yet theoretically unbounded. The practical
estimation is possible if round-trip times are sufficiently short in comparison to required
accuracy.

Principle:
 The UTC-synchronized time server S is used here.
 A process P sends requests to S and measures round-trip time Tround .
 In LAN, Tround should be around 1-10 ms .
 During this time, a clock with a 10-6 sec/sec drift rate varies by at most 10-8 sec.
 Hence the estimate of Tround is reasonably accurate .
 Now set clock to t + ½ Tround.
 Distributed Systems 4.4

Fig 4.1 Cristian’s algorithm

As per the given algorithm, the following results were found:
 Earliest time that S can have sent reply: t0 + min
 Latest time that S can have sent reply: t0 + Tround – min
 Total time range for answer: Tround - 2 * min
 Accuracy is ± (½Tround - min)

Problems in this system:

 Timer must never run backward.
 Variable delays in message passing / delivery occurs.
4.1.4 Berkeley algorithm
 Berkeley algorithm was developed to solve the problems of Cristian‟s algorithm.
 This algorithm does not need external synchronization.
 Master slave approach is used here.
 The master polls the slaves periodically about their clock readings.
 Estimate of local clock times is calculated using round trip.
 The average values are obtained from a group of processes.
 This method cancels out individual clock‟s tendencies to run fast and tells slave
processes by which amount of time to adjust local clock
 In case of master failure, master election algorithm is used.
 Distributed Systems 4.5

Fig 4.2: Berkeley Algorithm

In other words, the working of the algorithm can be summarized as,

 The time daemon asks all the other machines for their clock values.

 The machines answer the request.

 The time daemon tells everyone how to adjust their clock.

4.1.5 Network Time Protocol (NTP)

The Network Time Protocol defines architecture for a time service and a protocol to
distribute time information over the Internet.

Features of NTP:

 To provide a service enabling clients across the Internet to be synchronized

accurately to UTC.

 To provide a reliable service that can survive lengthy losses of connectivity.

 To enable clients to resynchronize sufficiently frequently to offset the rates of drift

found in most computers.

 To provide protection against interference with the time service, whether malicious
or accidental.
 Distributed Systems 4.6

Fig 4.3: NTP strata

 The NTP service is provided b y a network of servers located across the Internet.
 Primary servers are connected directly to a time source such as a radio clock
receiving UTC.
 Secondary servers are synchronized, with primary servers.
 The servers are connected in a logical hierarchy called a synchronization subnet.
 Arrows denote synchronization control, numbers denote strata. The levels are
called strata.
Working:
 NTP follows a layered client-server architecture, based on UDP message passing.

 Synchronization is done at clients with higher strata number and is less accurate
due to increased latency to strata 1 time server.
 If a strata 1 server fails, it may become a strata 2 server that is being synchronized
though another strata 1 server.

Modes of NTP:
NTP works in the following modes:

 Multicast:
 One computer periodically multicasts time info to all other computers on
network.

 These adjust clock assuming a very small transmission delay.

 Only suitable for high speed LANs; yields low but usually acceptable
synchronization.

 Procedure-call:
 This is similar to Christian‟s protocol .
 Distributed Systems 4.7

 Here the server accepts requests from clients.

 This is applicable where higher accuracy is needed, or where multicast is not

supported by the network‟s hardware and software

 Symmetric:

 This is used where high accuracy is needed.

Working of Procedure call and symmetric modes:

 All messages carry timing history information.

 The history includes the local timestamps of send and receive of the previous NTP
message and the local timestamp of send of this message

 For each pair i of messages (m, m‟) exchanged between two servers the following
values are being computed

- offsetoi : estimate for the actual offset between two clocks

-delay di : true total transmission time for the pair of messages.

Fig 4.4: Message exchange between NTP

Delay and offset:

 Let o be the true offset of B‟s clock relative to A‟s clock, and let t and t‟ the
true transmission times of m and m‟ (Ti , Ti-1 ... are not true time)

 The delay Ti-2 = Ti-3 + t + o..................(1)

Ti = Ti-1 + t‟ – o........................(2)

which leads to di = t + t‟ = Ti-2 - Ti-3 + Ti - Ti-1 (clock errors zeroed out à

(almost) true d)

 Offset oi = ½ (Ti-2 – Ti-3 + Ti-1 – Ti ) (only an estimate)

 Distributed Systems 4.8

Implementing NTP
 Statistical algorithms based on 8 most recent pairs are used in NTP to determine
quality of estimates.
 The value of oi that corresponds to the minimum di is chosen as an estimate for o .
 Time server communicates with multiple peers, eliminates peers with unreliable
data, favors peers with higher strata number (e.g., for primary synchronization
partner selection).
 NTP phase lock loop model: modify local clock in accordance with observed drift
rate.
 The experiments achieve synchronization accuracies of 10 msecs over Internet,
and 1 msec on LAN using NTP

4.2 LOGICAL TIME AND LOGICAL CLOCKS

 Synchronization between two processes without any form of interaction is not
needed.
 But when processes interact, they must be event ordered. This is done by logical
clocks.
 Consider a distributed system with n processes, p1, p2, …pn.
 Each process pi executes on a separate processor without any memory sharing.
 Each pi has a state si. The process execution is a sequence of events.
 As a result of the events, changes occur in the local state of the process.
 They either send or receive messages.
4.2.1 Lamport Ordering of Events
The partial ordering obtained by generalizing the relationship between two process is
called as happened-before relation or causal ordering or potential causal ordering. This
term was coined by Lamport. Happens-before defines a partial order of events in a distributed
system. Some events can‟t be placed in the order.
 We say e →i e‟ if e happens before e‟ at process i.
 e → e‟ is defined using the following rules:
 Local ordering: e → e‟ if e →ie‟ for any process i
 Messages: send(m) → receive(m) for any message m
 Transitivity: e → e‟‟ if e → e‟ and e‟ → e‟‟
 Distributed Systems 4.9

4.2.2 Logical Clocks

A Lamport logical clock is a monotonically increasing software counter, whose

value need bear no particular relationship to any physical clock.

 Lamport clock L orders events consistent with logical happens before

ordering. If e → e‟, then L(e) < L(e‟)
 But the converse is not true.
L(e) < L(e‟) does not imply e →
 e‟ Similar rules for concurrency are:
L(e) = L(e‟) implies e║e‟ (for distinct
e,e‟) e║e‟ does not imply L(e) = L(e‟)
This implies that Lamport clocks arbitrarily order some concurrent
events.

Fig 4.5: Lamport’s timestamps for the three process

Lamport’s Algorithm
 Assume that each process i keeps a local clock, Li. There are three rules:
1. At process i, increment Li before each event.
2. To send a message m at process i, apply rule 1 and then include the current local
time in the message: i.e., send(m,Li).
3. To receive a message (m,t) at process j, set Lj = max(Lj,t) and then apply rule 1
before time-stamping the receive event.
 The global time L(e) of an event e is just its local time. For an event e at process i,
L(e) = Li(e).
 Distributed Systems 4.10

Totally ordered logical clocks

• Many systems require a total-ordering of events, not a partial-ordering.
• Use Lamport‟s algorithm, but break ties using the process ID.

• L(e) = M * Li(e) +
i Where M = maximum number of processes

Vector clocks
The main aim of vector clocks is to order in such a way to match causality. The
expression, V(e) < V(e‟) if and only if e → e‟, holds for vector clocks, where V(e) is the
vector clock for event e. For implementing vector clock, label each event by vector V(e) [c1,
c2 …, cn]. ciis the events in process i that causally precede e.
• Each processor keeps a vector of values, instead of a single value.
• VCi is the clock at process i; it has a component for each process in the
system. VCi[i] corresponds to Pi„s local “time”.
VCi[j] represents Pi„s knowledge of the “time” at Pj(the # of events that Pi knows
have occurred at Pj
• Each processor knows its own “time” exactly, and updates the values of other
processors‟ clocks based on timestamps received in messages.

Fig 4.6: Vector Timestamps

Vector timestamps have the disadvantage, compared with Lamport timestamps, of
taking up an amount of storage and message payload that is proportional to N, the
number of processes.

4.3 GLOBAL STATES

This section checks whether a property is true in a distributed systems. This is checked
for the following
problems:
 Distributed Systems 4.11

 distributed garbage collection

 deadlock detection
 termination detection

 Debugging

The global state of a distributed system consists of the local state of each process,
together with the messages that are currently in transit, that is, that have been
sent but not delivered.

Distributed Garbage Collection

 An object is considered to be garbage if there are no longer any references to it
anywhere in the distributed system.

Fig 4. 7: Distributed Garbage Collection

 To check that an object is garbage, verify that there are no references to it anywhere.
 The process p1 has two objects that both have references :
 One has a reference within p1 itself
 p2 has a reference to the other.
 Process p2 has one garbage object, with no references to it anywhere in the system.
 It also has an object for which neither p1 nor p2 has a reference, but there is a
reference to it in a message that is in transit between the processes.
Distributed deadlock detection:
 A distributed deadlock occurs when each of a collection of processes waits for another
process to send it a message, and where there is a cycle in the graph of this „waits-for‟
relationship.
 In the following figure, processes p1 and p2 are each waiting for a message from the
other, so this system will never make progress.
4.12 Synchronization and replicatio Distributed Systems 4.12
n
Fig 4.8: Distributed deadlock detection
Termination detection
 Detecting termination is a problem is to test whether each process has halted.
 Find whether a process is either active or passive.
 A passive process is not engaged in any activity of its own but is prepared to respond
with a value requested by the other.
 The phenomena of termination and deadlock are similar:
 A deadlock may affect only a subset of the processes in a system, whereas all
processes must have terminated.
 The pro cess passivity is not the same as waiting in a deadlock cycle: a
deadlocked process is attempting to perform a further action, for which
another process waits; a passive process is not engaged in any activity.

Fig 4.9: Termination of process

Distributed debugging:
Consider an application with process pi and a variable x, where i=1, 2, …, N. As the
program executes the variables may change the value. The value must be within the range.
The relationships between variables must be calculated only for the variables which executes
at same time.

4.3.1 Global States and Consistency cuts (Distributed Snapshots)

Distributed Snapshot represents a state in which the distributed system might have
been in. A snapshot of the system is a single configuration of the system.
 Distributed Systems 4.13

A distributed snapshot should reflect a consistent state. A global state is consistent if it

could have been observed by an external observer. For a successful Global State, all states
must be consistent
 If we have recorded that a process P has received a message from a process Q, then
we should have also recorded that process Q had actually send that message.
 Otherwise, a snapshot will contain the recording of messages that have been
received but never sent.
 The reverse condition (Q has sent a message that P has not received) is allowed.
The notion of a global state can be graphically represented by what is called a cut. A
cut represents the last event that has been recorded for each process.
The history of each process if given by:
history(p i )  hi  ei 0 , ie1, ie2 ,...

Each event either is an internal action of the process. We denote by s i kt he state of

process pi immediately before the kth event occurs. The state siin the global state S
correspondin g to the cut C is that of pi immediately after the last event processed by p i in the
cut – eici. The set of events eici is called the frontier of the cut.

Fig 4.10: Types of cuts

In the above figure, the <e11, e 20 >and <e 22 , e13 > are the frontiers.The <e 11 , e02 > is
inconsistent because at p2 it includes the receipt of the message m1 , but at p1 it does not
include the sending of that message. The frontier <e22, e13> is consistent since it includes
both the sending and the receipt of message m1 and the sending but not the receipt of message
m2.
A consistent global state is one that corresponds to a consistent cut. We may
characterize the e execution of a distributed s yste m as a series of transitions between global
states of the system:
S 0 S 1 S 2  …
A run is a total ordering of all the events in a global history that is consistent with
each local history‟s ordering. A linearization or consistent run is an ordering of the events
in a global history that is consistent with this happened-before relation o on H. A linearization
is also a
run.
 Distributed Systems 4.14

4.3.2 Global state predicates, stability, safety and liveness

 Detecting a condition such as deadlock or termination amounts to evaluating a
global state predicate.
 A global state predicate is a function that maps from the set of global states of
processes in the system
 One of the useful characteristics of the predicates associated with the state of an
object being garbage, of the system being deadlocked or the system being
terminated is that they are all stable: once the system enters a state in which the
predicate is True, it remains True in all future states reachable
from that state.
 By contrast, when we monitor or debug an application we are often interested in
non-stable predicates, such as that in our example of variables whose difference is
supposed to be bounded.
4.3.3 Snapshot algorithm of Chandy and
Lamport Assumptions of the algorithm
 The algorithm to determine global states records process states locally and the
states are collected by a designated server.
 No failures in c hannels and processes – exactly once y.
 Unidirectional deliver channel and FIFO ordered message
 There is always delivery. a path between two processes.
 Global snapsho t initiation at any process at any time.
 No process acti vity that halt during snapshot.

Fig 4.11: Process and organization in distributed systems

Algorithm:
 If a process Q receives the marker requesting a snapshot for the first time, it
considers the process that sent the marker as its predecessor.
 Distributed Systems 4.15

 When Q completes its part of the snapshot, it sends its predecessor a DONE
message.

 By recursion, when the initiator of the distributed snapshot has received a DONE
message from all its successors, it knows that the snapshot has been completely
taken.

4.4 COORDINATION AND AGREEMENT

When more that one process run in an execution environment, they need to coordinate
their actions. To solve problems in coordination avoid fixed master-salve relationship to avoid
single points of failure for fixed master. Distributed mutual exclusion for resource sharing can
also be used.

When a collection of process share resources, mutual exclusion is needed to prevent

interference and ensure consistency. In this case, there is no need for shared variables or
facilities are provided by single local kernel to solve it. This solution is based on message
passing. The main issue to be addressed while implementing the above method is the failure.

4.4.1 Failure Assumptions and failure Detectors

 The message passing paradigm is based on reliable communication channels.

 But there can be process failures (i.e.) the whole process crashes.

 The failure can be detected when the object/code in a process that detects failures
of other processes.

 There are two types of failure detectors: unreliable failure detector and reliable
failure detector.

 The following are features of the unreliable failure detectors:

• unsuspected or suspected (i.e.) there can be no evidence of failure

• each process sends ``alive'' message to everyone else

• not receiving ``alive'' message after timeout

• This is present in most practical systems

 The following are features of the reliable failure detectors:

• unsuspected or failure

• They are present in synchronous system

 Distributed Systems 4.16

4.5 DISTRIBUTED MUTUAL EXCLUSION

The mutual exclusion makes sure that concurrent process access shared resources or
data in a serialized way. If a process, say Pi , is executing in its critical section, then no other
processes can be executing in their critical sections.

Distributed mutual exclusion provide critical region in a distributed environment.

4.5.1 Algorithms for Distributed Mutual Exclusion

Consider there are N processes and the processes do not fail. The basic assumption is
the message delivery system is reliable. So the methods for the critical region are:
 enter() : Enter the critical section block if necessary
 resourceAccesses():Access the shared resources

 exit(): Leaves the critical section. Now other processes can enter critical
section. The following are the requirements for Mutual Exclusion (ME):
 [ME1] safety: only one process at a time

 [ME2] liveness: eventually enter or exit

 [ME3] happened-before ordering: ordering of enter() is the same as HB
ordering
The second requirement implies freedom from both deadlock and starvation.
Starvation involves fairness condition.

Performance Evaluation:
The following are the criteria for performance measures:
 Bandwidth consumption, which is proportional to the number of messages sent
in each entry and exit operations.
 The client delay incurred by a process at each entry and exit operation.
 Throughput of the system: Rate at which the collection of processes as a whole
can access the critical section.

4.5.2 Central Sever Algorithm

 This employs the simplest way to grant permission to enter the critical section by
using a server.
 A process sends a request message to server and awaits a reply from it.
 Distributed Systems 4.17

 If a reply constitutes a token signifying the permission to enter the critical section.
 If no other process has the token at the time of the request, then the server replied
immediately with the token.
 If token is currently held by another process, then the server does not reply but
queues the request.
 Client on exiting the critical section, a message is sent to server, giving it back the
token.

Fig 4. 12: Central Server Algorithm

The central server algorithm fulfils ME1 and ME2 but not ME3 (i.e.) safety and
liveness is ensured but ordering is not satisfied. Also the performance of the algorithm is
measured as follows:
 Bandwidth: This is measured by entering and exiting messages. Entering takes
two messages ( request followed by a grant) which are delayed by the round-
trip time. Exiting takes one release message, and does not delay the exiting
process.
 Throughput is measured by synchronization delay, round-trip of a release
message and grant message.

4.5.3 Ring Based Algorithm

 This provides a simplest way to arrange mutual exclusion between N processes
without requiring an additional process is arrange them in a logical ring.
 Distributed Systems 4.18

 Each process pi has a communication channel to the next process in the ring as
follows: p(i+1)/mod N.
 The unique token is in the form of a message passed from process to process in a
single direction clockwise.
 If a process does not require to enter the CS when it receives the token, then it
immediately forwards the token to its neighbor.
 A process requires the token waits until it receives it, but retains it.
 To exit the critical section, the process sends the token on to its neighbor.

Fig 4.13: Ring based algorithm

This algorithm satisfies ME1 and ME2 but not ME (i.e.) safety and liveness are
satisfied but not ordering. The performance measures include:
 Bandwidth: continuously consumes the bandwidth except when a process is
inside the CS. Exit only requires one message.
 Delay: experienced by process is zero message(just received token) to N
messages(just pass the token).
 Throughput: synchronization delay between one exit and next entry is anywhere
from 1(next one) to N (self) message transmission.
4.5.4 Multicast Synchronisation
 This exploits mutual exclusion between N peer processes based upon multicast.
 Processes that require entry to a critical section multicast a request message, and can
enter it only when all the other processes have replied to this message.
 Distributed Systems 4.19
 The condition under which a process replies to a request are designed to ensure ME1
ME2 and ME3 are met.
 Each process pi keeps a Lamport clock. Message requesting entry are of the form<T,
pi>.
 Each process records its state of either RELEASE, WANTED or HELD in a variable
state.
 If a process requests entry and all other processes is RELEASED, then all
processes reply immediately.
 If some process is in state HELD, then that process will not reply until it is
finished.
 If some process is in state WANTED and has a smaller timestamp than the
incoming request, it will queue the request until it is finished.
 If two or more processes request entry at the same time, then whichever bears the
lowest timestamp will be the first to collect N-1 replies.

Fig 4.13: Multicast Synchronisation

 In the above figure, P1 and P2 request CS concurrently.
 The timestamp of P1 is 41 and for P2 is 34.
 When P3 receives their requests, it replies immediately.
 When P2 receives P1‟s request, it finds its own request has the lower timestamp, and
so does not reply, holding P1 request in queue.
 However, P1 will reply. P2 will enter CS. After P2 finishes, P2 reply P1 and P1 will
enter CS.
 Granting entry takes 2(N-1) messages, N-1 to multicast request and N-1 replies.
 Distributed Systems 4.20

Performance Evaluation:
 Bandwidth consumption is high.
 Client delay is again 1 round trip time
 Synchronization delay is one message transmission time.

4.5.5 Maekawa’s Voting Algorithm

• In this algorithm, it is not necessary for all of its peers to grant access. Only need
to obtain permission to enter from subsets of their peers, as long as the subsets
used by any two processes overlap.
• Think of processes as voting for one another to enter the CS. A candidate
process must collect sufficient votes to enter.
• Processes in the intersection of two sets of voters ensure the safety property ME1
by casting their votes for only one candidate.
• A voting set Vi associated with each process pi.
• There is at least one common member of any two voting sets, the size of all
voting set are the same size to be fair.

• The optimal solution to minimizes K is K~sqrt(N) and M=K.

• The algorithm is summarized as

follows: Vi  {P1, P2, … PN}

Such that for all i, j = 1, 2, … N
Pi  Vi
Vi  Vj  
|Vi | = K
Each process is contained in M of the voting set Vi
Maekawa’s Voting Algorithm
On initialization
state := RELEASED;
voted := FALSE;
For pito enter the critical section
state := WANTED;
 Distributed Systems 4.21

Multicast request to all processes in Vi;

Wait until (number of replies received = K);
state := HELD;
On receiptof a request frompi at pj
if (state = HELD orvoted = TRUE)
then
queue request from pi without replying;
else
send reply to pi;
voted := TRUE;
end if
For pi to exit the critical section
state := RELEASED;
Multicast release to all processes in Vi;
On receiptof a release frompi at pj
if (queue of requests is non-empty)
then
remove head of queue – from pk, say;
send reply to pk;
voted := TRUE;
else
voted := FALSE;
end if

 The ME1 is met.

 If two processes can enter CS at the same time, the processes in the intersection of
two voting sets would have to vote for both.
 The algorithm will only allow a process to make at most one vote between successive
receipts of a release message.
 This algorithm is deadlock prone.
 Distributed Systems 4.22

 If three processes concurrently request entry to the CS, then it is possible for p1 to
reply to itself and hold off p2; for p2 rely to itself and hold off p3; for p3 to reply to
itself and hold off p1.

 Each process has received one out of two replies, and none can proceed.

 If process queues outstanding request in happen-before order, ME3 can be satisfied

and will be deadlock free.

Performance Evaluation:

 Bandwidth utilization is 2sqrt(N) messages per entry to CS and sqrt(N) per exit.

 Client delay is the same as Ricart and Agrawala‟salgorithm, one round-trip time.

 Synchronization delay is one round-trip time which is worse than R&A

4.5.6 Fault Tolerance

 The reactions of the algorithms when messages are lost or when a process crashes
is fault tolerance.

 None of the algorithm that we have described would tolerate the loss of messages
if the channels were unreliable.

 The ring-based algorithm cannot tolerate any single process crash failure.

 Maekawa‟salgorithm can tolerate some process crash failures: if a crashed

process is not in a voting set that is required.

 The central server algorithm can tolerate the crash failure of a client process that
neither holds nor has requested the token.

4.6 ELECTIONS

An algorithm for choosing a unique process to play a particular role is called an

election algorithm.

The requirements for election algorithms:

 E1(safety): a participant pi haselectedi =  or elected P where  indicates a
value that is not defined.
 E2(liveness): All processes Pi participate in election process and eventually
setelectedi!= or crash.
 Distributed Systems 4.23

4.6.1 Ring based Election Algorithm

 All the processes arranged in a logical ring.
 Each process has a communication channel to the next process.
 All messages are sent clockwise around the ring.
 Assume that no failures occur, and system is asynchronous.
 The ultimate goal is to elect a single process coordinator which has the largest
identifier.

Fig 4.14: Election process using Ring based election algorithm

Steps in election process:
1. Initially, every process is marked as non-participant. Any process can begin an
election.
2. The starting processes marks itself as participant and place its identifier in a message
to its neighbour.
3. A process receives a message and compares it with its own. If the arrived identifier is
larger, it passes on the message.
4. If arrived identifier is smaller and receiver is not a participant, substitute its own
identifier in the message and forward if. It does not forward the message if it is already
a participant.
5. On forwarding of any case, the process marks itself as a participant.
 Distributed Systems 4.24

6. If the received identifier is that of the receiver itself, then this process‟ s identifier
must be the greatest, and it becomes the coordinator.
7. The coordinator marks itself as non-participant set elected_i and sends an elected
message to its neighbour enclosing its ID.
8. When a process receives elected message, marks itself as a non-participant, sets its
variable elected_i and forwards the message.
The election was started by process 17. The highest process identifier encountered so
far is 24.

Requirements:
 E1 is met. All identifiers are compared, since a process must receive its own ID
back before sending an elected message.
 E2 is also met due to the guaranteed traversals of the ring.

 Tolerates no failure makes ring algorithm of limited practical use.

Performance Evaluation
 If only a single process starts an election, the worst-performance case is then the
anti-clockwise neighbour has the highest identifier. A total of N-1 messages is
used to reach this neighbour. Then further N messages are required to announce its
election. The elected message is sent N times. Making 3N-1 messages in all.
 Turnaround time is also 3N-1 sequential message transmission time.

4.6.2 Bully Algorithm

 This algorithm allows process to crash during an election, although it assumes the
message delivery between processes is reliable.
 Assume that the system is synchronous to use timeouts to detect a process failure
and each process knows which processes have higher identifiers and that it can
communicate with all such processes.
 In this algorithm, there are three types of messages:

o Election message: This is sent to announce an election message. A process

begins an election when it notices, through timeouts, that the coordinator
has failed. T=2Ttrans+Tprocess From the time of sending
o Answermessage: This is sent in response to an election message.
o Coordinatormessage: This is sent to announce the identity of the elected
process.
 Distributed Systems 4.25

Fig 4.15: Stages in Bully Algorithm

Election process:
 The process begins a election by sending an election message to these processes that
have a higher ID and awaits an answer in response.
 If none arrives within time T, the process considers itself the coordinator and sends
coordinator message to all processes with lower identifiers.
 Otherwise, it waits a further time T‟ for coordinator message to arrive. If none, begins
another election.
 If a process receives a coordinator message, it sets its variable elected_i to be the
coordinator ID.
 If a process receives an election message, it send back an answer message and begins
another election unless it has begun one already.
Requirements:
 E1 may be broken if timeout is not accurate or replacement.
 Suppose P3 crashes and replaced by another process. P2 set P3 as coordinator and P1
set P2 as coordinator.
 E2 is clearly met by the assumption of reliable transmission.
 Distributed Systems 4.26

Performance Evaluation
 Best case the process with the second highest ID notices the coordinator‟s failure.
Then it can immediately elect itself and send N-2 coordinator messages.
 The bully algorithm requires O(N^2) messages in the worst case - that is, when the
process with the least ID first detects the coordinator‟s failure. For then N-1 processes
altogether begin election, each sending messages to processes with higher ID.

4.7 TRANSACTION AND CONCURRENCY CONTROL

Fundamentals of transactions and concurrency control
 A transaction is generally atomic.
 The state of the transaction being done is not visible. If it is not done completely,
any changes it made will be undone. This is known as rollback.
 Concurrency is needed when multiple users want to access the same data at the
same time.
 Concurrency control (CC) ensures that correct results for parallel operations are
generated.
 CC provides rules, methods, design methodologies and theories to maintain the
consistency of components operating simultaneously while interacting with the
same object.

A Transaction defines a sequence of server operations that is guaranteed to

be atomic in the presence of multiple clients and server crash.

All concurrency control protocols are based on serial equivalence and are derived from
rules of conflicting operations:
 Locks used to order transactions that access the same object according to request
order.
 Optimistic concurrency control allows transactions to proceed until they are ready
to commit, whereupon a check is made to see any conflicting operation on objects.
 Timestamp ordering uses timestamps to order transactions that access the same
object according to their starting time.

Synchronisation without transactions

 Synchronization without transactions are done with multiple threads.

 The use of multiple threads is beneficial to the performance. Multiple threads may
access the same objects.
 Distributed Systems 4.27

 In Java, Synchronized keyword can be applied to method so only one thread at a

time can access an object.
 If one thread invokes a synchronized method on an object, then that object is
locked, another thread that invokes one of the synchronized method will be
blocked.

Enhancing Client Cooperation by Signaling

 The clients may use a server as a means of sharing some resources.
 In some applications, threads need to communicate and coordinate their actions by
signaling.

Failure modes in transactions

 The following are the failure modes: Writes to permanent storage may fail, either
by writing nothing or by writing wrong values.
 Servers may crash occasionally.
 There may be an arbitrary delay before a message arrives.

4.8 TRANSACTIONS
 Transaction originally from database management systems.
 Clients require a sequence of separate requests to a server to be atomic in the sense
that:
• They are free from interference by operations being performed on behalf of
other concurrent clients.
• Either all of the operations must be completed successfully or they must have
no effect at all in the presence of server crashes.
Atomicity
 All or nothing: a transaction either completes successfully, and effects of all of its
operations are recorded in the object, or it has no effect at all.
 Failure atomicity: effects are atomic even when server crashes
 Durability: after a transaction has completed successfully, all its
effects are saved in permanent storage for recover later.

Isolation
Each transaction must be performed without interference from other transactions. The
intermediate effects of a transaction must not be visible to other transactions.
 Distributed Systems 4.28

4.8.1 Concurrency
Control Serial Equivalence
 If these transactions are done one at a time in some order, then the final result will
be correct.
 If we do not want to sacrifice the concurrency, an interleaving of the operations of
transactions may lead to the same effect as if the transactions had been performed
one at a time in some order.
 We say it is a serially equivalent interleaving.
 The use of serial equivalence is a criterion for correct concurrent execution to
prevent lost updates and inconsistent retrievals.
Conflicting Operations
 When we say a pair of operations conflicts we mean that their combined effect
depends on the order in which they are executed. E.g. read and write
 There are three ways to ensure serializability:
 Locking
 Timestamp ordering
 Optimistic concurrency control

Process 1 Process 2 Conflict Reason

Read Read No -
Read Write Yes The result of these operations dependon their
order of execution.
Write Write Yes The result of these operations depend on their
order of execution.

4.9 NESTED TRANSACTIONS

 Nested transactions extend the transaction model by allowing transactions to be
composed of other transactions.
 Thus several transactions may be started from within a transaction, allowing
transactions to be regarded as modules that can be composed as required.

 The outermost transaction in a set of nested transactions is called the top-level

transaction.
 Transactions other than the top-level transaction are called sub-transactions.
 Distributed Systems 4.29

Fig 4.16: Nested Transactions

 T is a top-level transaction that starts a pair of sub-transactions, T1 and T2.

 The sub-transaction T1 starts its own pair of sub-transactions, T11 and T22.

 Also, sub-transaction T2 starts its own sub-transaction, T21, which starts

another sub-transaction, T211.

 A sub-transaction appears to be atomic to its parent with respect to transaction

failures
and to concurrent access.

 Sub-transactions at the same level can run concurrently, but their access to
common objects is serialized.

 Each sub-transaction can fail independently of its parent and of the other sub-
transactions.

 When a sub-transaction aborts, the parent transaction can choose an alternative

sub-transaction to complete its task.

 If one or more of the sub-transactions fails, the parent transaction could record
the fact and then commit, with the result that all the successful child
transactions commit.

 It could then start another transaction to attempt to redeliver the messages that
were not sent the first time.

Advantages of Nested Transactions

 Sub- transactions at same level can run concurrently.

 Sub- transactions can commit or abort independently.

 Distributed Systems 4.30

The rules for committing of nested transactions are:

 A transaction may commit or abort only after its child transactions have
completed.

 When a sub-transaction completes, it makes an independent decision either to

commit provisionally or to abort. Its decision to abort is final.

 When a parent aborts, all of its sub-transactions are aborted.

 When a sub -transaction aborts, the parent can decide whether to abort or not.

 If the top-level transaction commits, then all of the sub-transactions that have
provisionally committed can commit too, provided that none of their ancestors has
aborted.

4.10 LOCKS
 A simple example of a serializing mechanism is the use of exclusive locks.

 Server can lock any object that is about to be used by a client.

 If another client wants to access the same object, it has to wait until the object is
unlocked in the end.

 There are two types of locks: Read and Write.

 Read locks are shared locks (i.e.) they do not bring about conflict.

 Write locks are exclusive locks, since the order in which write is done may give
rise to a conflict.
Lock Read conflict Write conflict
None No No
Read No Yes
Write Yes Yes

 An object can be read and write.

 From the compatibility table, we know pairs of read operations from different
transactions do not conflict.
 So a simple exclusive lock used for both read and write reduces concurrency more
than necessary. (Many readers/Single writer).
 Distributed Systems 4.31

The following rules could be framed:

1. If T has already performed a read operation, then a concurrent transaction U
must not write until T commits or aborts.
2. If T already performed a write operation, then concurrent U must not read or
write until T commits or aborts.
Lock implementation
• The granting of locks will be implemented by a separate object in the server called
the lock manager.
• The lock manager holds a set of locks, for example in a hash table.
• Each lock is an instance of the class Lock and is associated with a particular
object.
• Each instance of Lock maintains the following information in its instance
variables:
 the identifier of the locked object
 the transaction identifiers of the transactions that currently hold the lock
 a lock type
Two phase Locking
The basic two-phase locking (2PL) protocol states:
 A transaction T must hold a lock on an item x in the appropriate mode before T
accesses x.
 If a conflicting lock on x is being held by another transaction, T waits.
 Once T releases a lock, it cannot obtain any other lock
subsequently. In two phase locking, a transaction is divided into two
phases:
 A growing phase (obtaining locks)
 A shrinking phase (releasing locks)
This lock ensures conflict serializability.Lock-point is the point where the
transaction obtains all the locks.With 2PL, a schedule is conflict equivalent to aa
serial schedule ordered by the lock-point of the transactions.
Strict Two Phase Locking
 Because transaction may abort, strict execution are needed to prevent dirty reads
and premature writes, which are caused by read or write to same object accessed
by another earlier unsuccessful transaction that already performed an write
operation.
 Distributed Systems 4.32

 So to prevent this problem, a transaction that needs to read or write an object must
be delayed until other transactions that wrote the same object have committed or
aborted.
 The rule in Strict Two Phase Locking:
 Any locks applied during the progress of a transaction are helduntil the
transaction commits or aborts.

Fig 4. 16: Strict two phase locking between T1 and T2

1. When an operation accesses an object within a transaction:
(a) If the object is not already locked, it is locked and the operation proceeds.
(b) If the object has a conflicting lock set by another transaction, the transaction
must wait until it is unlocked.
(c) If the object has a non-conflicting lock set by another transaction, the lock is
shared and the operation proceeds.
(d) If the object has already been locked in the same transaction, the lock will be
promoted if necessary and the operation proceeds. (Where promotion is
prevented by a conflicting lock, rule (b) is used.)
2. When a transaction is committed or aborted, the server unlocks all objects it locked for
the transaction
A transaction with a read lock that is shared by other transactions cannot promote its
read lock to a write lock, because write lock will conflict with other read locks.

4.10.1 Deadlocks

Deadlock is a state in which each member of a group of transactions is waiting for

some other member to release a lock.
 Distributed Systems 4.33

 A wait-for graph can be used to represent the waiting relationships between current
transactions.
 In a wait-for graph the nodes represent transactions and the edges represent wait-for
relationships between transactions – there is an edge from node T to node U when
transaction T is waiting for transaction U to release a lock.

Fig 4.17: Wait for

graph
Fig 4.18: A cycle in wait for graph
Deadlock prevention:
 Simple way is to lock all of the objects used by a transaction when it starts.
 It should be done as an atomic action to prevent deadlock. a. inefficient, say lock
an object you only need for short period of time. b.

 Hard to predict what objects a transaction will require.

 Judge if system can remain in a Safe state by satisfying a certain resource request.
Banker‟s algorithm.
 Order the objects in certain order.
 Acquiring the locks need to follow this certain order.
 Distributed Systems 4.34

Safe State
 System is in safe state if there exists a sequence <P 1, P2, …, Pn> of ALL the
processes is the systems such that for each P i, the resources that Pi can still
request can be satisfied by currently available resources + resources held by all the
Pj, with j <i.
 If a system is in safe state, then there are no deadlocks.
 If a system is in unsafe state, then there is possibility of deadlock.
 Avoidance is to ensure that a system will never enter an unsafe state.
Deadlock Detection
• Deadlock may be detected by finding cycles in the wait-for-graph. Having detected
a deadlock, a transaction must be selected for abortion to break the cycle.
o If lock manager blocks a request, an edge can be added. Cycle should be
checked each time a new edge is added.
o One transaction will be selected to abort in case of cycle. Age of
transaction and number of cycles involved when selecting a victim
• Timeouts is commonly used to resolve deadlock. Each lock is given a limited
period in which it is invulnerable. After this time, a lock becomes vulnerable.
 If no other transaction is competing for the object, vulnerable object remained
locked. However, if another transaction is waiting, the lock is broken.
Disadvantages:
 Transaction aborted simply due to timeout and waiting transaction even if there is
no deadlock.
 Hard to set the timeout time

4.11 OPTIMISTIC COCURRENCY CONTROL

The locking and serialization of transaction has numerous disadvantages.:
 Lock maintenance represents an overhead that is not present in systems that do not
support concurrent access to shared data. Locking sometimes are only needed for
some cases with low probabilities.
 The use of lock can result in deadlock. Deadlock prevention reduces concurrency
severely. The use of timeout and deadlock detection is not ideal for interactive
programs.
 To avoid cascading aborts, locks cannot be released until the end of the
transaction. This may reduce the potential for concurrency.
 Distributed Systems 4.35
 It is based on observation that, in most applications, the likelihood of two clients‟
transactions accessing the same object is low.
 Transactions are allowed to proceed as though there were no possibility of conflict
with other transactions until the client completes its task and issues a
closeTransaction request.
 When conflict arises, some transaction is generally aborted and will need to be
restarted by the client.
 Each transaction has the following phases:
 Working phase:
 Each transaction has a tentative version of each of the objects that it
updates.
 This is a copy of the most recently committed version of the object.
 The tentative version allows the transaction to abort with no effect on the
object, either during the working phase or if it fails validation due to other
conflicting transaction.
 Several different tentative values of the same object may coexist.
 In addition, two records are kept of the objects accessed within a
transaction, a read set and a write set containing all objects either read or
written by this transaction.
 Read are performed on committed version (no dirty read can occur) and
write record the new values of the object as tentative values which are
invisible to other transactions.
 Validation phase:
 When closeTransaction request is received, the transaction is validated to
establish whether or not its operations on objects conflict with operations
of other transaction on the same objects.
 If successful, then the transaction can commit.
 If fails, then either the current transaction or those with which it conflicts
will need to be aborted.
 Update phase:
 If a transaction is validated, all of the changes recorded in its tentative
versions are made permanent.
 Read-only transaction can commit immediately after passing validation.
 Write transactions are ready to commit once the tentative versions have
been recorded in permanent storage.
 Distributed Systems 4.36

Validation of Transactions
 Validation uses the read-write conflict rules to ensure that the scheduling of a
particular transaction is serially equivalent with respect to all other overlapping
transactions- that is, any transactions that had not yet committed at the time this
transaction started.
 Each transaction is assigned a number when it enters the validation phase (when
the client issues a closeTransaction).
 Such number defines its position in time.
 A transaction always finishes its working phase after all transactions with lower
numbers.
 That is, a transaction with the number Ti always precedes a transaction with
number Tj if i < j.
 The validation test on transaction Tv is based on conflicts between operations in
pairs of transaction Ti and Tv, for a transaction Tv to be serializable with respect
to an overlapping transaction Ti, their operations must conform to the below rules.

Rule No Tv T1 Rule
1 Write Read Ti must not read objects written by Tv
2 Read Write Tvmust not read objects written by Ti
3 Write Write Ti must not read objects written by Tvand Tvmust not
read objects written by Ti.

Types of Validation
 Backward Validation: checks the transaction undergoing validation with other
preceding overlapping transactions- those that entered the validation phase before it.
 Forward Validation: checks the transaction undergoing validation with other later
transactions, which are still active

Backward Validation Forward Validation

Backward validation of transaction Tv Forward validation of transaction Tv
boolean valid = true; boolean valid = true;
for (intTi = startTn+1; Ti<= for (intTid = active1; Tid<= activeN;
finishTn; Ti++){ Tid++){
if (read set of Tv intersects if (write set of Tv intersects
write set of Ti) valid = false; read set of Tid) valid = false;
} }
 Distributed Systems 4.37

Starvation
 When a transaction is aborted, it will normally be restarted by the client
program.
 There is no guarantee that a particular transaction will ever pass the validation
checks, for it may come into conflict with other transactions for the use of objects
each time it is restarted.

The prevention of a transaction ever being able to commit is called starvation.

4.12 TIMESTAMP ORDERING

 Timestamps may be assigned from the server‟s clock or a counter that is
incremented whenever a timestamp value is issued.
 Each object has a write timestamp and a set of tentative versions, each of which
has a write timestamp associated with it; and a set of read timestamps.
 The write timestamps of the committed object is earlier than that of any of its
tentative versions, and the set of read timestamps can be represented by its
maximum member.
 Whenever a transaction‟s write operation on an object is accepted, the server
creates a new tentative version of the object with write timestamp set to the
transaction timestamp. Whenever a read operation is accepted, the timestamp of
the transaction is added to its set of read timestamps.
 When a transaction is committed, the values of the tentative version become the
values of the object, and the timestamps of the tentative version become the write
timestamp of the corresponding object.
 Each request from a transaction is checked to see whether it conforms to the
operation conflict rules.
 Conflict may occur when previous done operation from other transaction Ti is later
than current transaction Tc. It means the request is submitted too late.
Rule Tc Ti Description
1 Write Read Tc must not write an object that has been read by any
Ti where Ti>Tcthis requires that Tc>= the maximum
read timestamp of the object.
2 Write Write Tc must not write an object that has been written by
any Ti where Ti>Tcthis requires that Tc> the write
timestamp of the committed object.
 Distributed Systems 4.38

3 Read Write Tc must not read an object that has been written by
any Ti where Ti>Tcthis requires that Tc>the write
timestamp of the committed object.

Timestamp ordering by read rule:

if ( Tc> write timestamp on committed version of D) {
let Dselected be the version of D with the maximum write timestamp ≤ Tc
if (Dselected is committed)
perform read operation on the version Dselected
else
Wait until the transaction that made version Dselected commits or aborts
then reapply the read rule
} else
Abort transaction Tc

Timestamp ordering by write rule

if (Tc ≥ maximum read timestamp on D&&
Tc> write timestamp on committed version of D)
perform write operation on tentative version of D with write timestamp Tc
else /* write is too late */
Abort transaction Tc

Multi-version timestamp ordering

 In multi-version timestamp ordering, a list of old committed versions as well as
tentative versions is kept for each object.
 This list represents the history of the values of the object.
 The benefit of using g multiple versions is that read operations that arrive too late
need not be rejected.
 Each version has a read timestamp recording the largest timestamp of any
transaction that has read from it in addition to a write timestamp.
 As before, whenever a write operation is accepted, it is directed to a tentative
version with the write timestamp of the transaction.
 Distributed Systems 4.39

 Whenever a read operation is carried out, it is directed to the version with the
largest write timestamp less than the transaction timestamp.
 If the transaction timestamp is larger than the read timestamp of the version being
used, the read timestamp of the version is set to the transaction timestamp.
 When a read arrives late, it can be allowed to read from an old committed version,
so there is no need to abort late read operations.
 In multi-version timestamp ordering, read operations are always permitted,
although they may have to wait for earlier transactions to complete, which ensures
that executions are recoverable.
 There is no conflict between write operations of different transactions, because
each transaction writes its own committed version of the objects it accesses.
 The rule in this ordering is

Tcmust not write objects that have been read by any Ti where Ti >Tc

Multi-version Timestamp ordering write rule

if (read timestamp of DmaxEarlier” Tc)
perform write operation on a tentative version of D with write timestamp Tc
else abort transaction Tc.

4.13 ATOMIC COMMIT PROTOCOL

 The atomicity property of transactions demands when a distributed transaction
comes to an end, either all of its operations are carried out or none of them.
 In the case of a distributed transaction, the client has requested operations at more
than one server.
 A transaction comes to an end when the client requests that it be committed or
aborted.
 A simple way to complete the transaction in an atomic manner is for the
coordinator to communicate the commit or abort request to all of the participants
in the transaction and to keep on repeating the request until all of them have
acknowledged that they have carried it out. This is one phase atomic commit
protocol.
 Distributed Systems 4.40

Two Phase Commit Protocol

 The two-phase commit protocol is designed to allow any participant to abort its
part of a transaction.
 Due to the requirement for atomicity, if one part of a transaction is
aborted, then the whole transaction must be aborted.
 In the first phase of the protocol, each participant votes for the transaction to be
committed or aborted.
 Once a participant has voted to commit a transaction, it is not allowed to abort it.
 In the second phase of the protocol, every participant in the transaction carries out
the joint decision.
 If any one participant votes to abort, then the decision must be to abort the
transaction.
 If all the participants vote to commit, then the decision is to commit the
transaction.
 The following are the operations in two phase commit protocol:
 canCommit?(trans)Yes / No: Call from coordinator to participant to ask
whether it can commit a transaction. Participant replies with its vote.
 doCommit(trans):Call from coordinator to participant to tell participant to
commit its part of a transaction.
 doAbort(trans): Call from coordinator to participant to tell participant to abort
its part of a transaction.
 haveCommitted(trans, participant): Call from participant to coordinator to
confirm that it has committed the transaction.
 getDecision(trans) Yes / No: Call from participant to coordinator to ask for
the decision on a transaction when it has voted Yes but has still had no reply
after some delay. Used to recover from server crash or delayed messages

Two phase commit protocol

Phase 1 (voting phase):

1. The coordinator sends a canCommit? request to each of the participants in the transaction.

2. When a participant receives a canCommit? request it replies with its vote (Yes or No) to
the coordinator. Before voting Yes, it prepares to commit by saving objects in permanent
storage. If the vote is No, the participant aborts immediately.
 Distributed Systems 4.41

Phase 2 (completion according to outcome of vote):

3. The coordinator collects the votes (including its own).
(a) If there are no failures and all the votes are Yes, the coordinator decides to commit the
transaction and sends a doCommitrequest to each of the participants.
(b) Otherwise, the coordinator decides to abort the transaction and sends doAbortrequests to
all participants that voted Yes.
4. Participants that voted Yes are waiting for a doCommitor doAbortrequest from the
coordinator. When a participant receives one of these messages it acts accordingly and, in the
case of commit, makes a have Committedcall as confirmation to the coordinator.

Fig 4.16: Two phase commit protocol

Two phase commit for nested transactions
 The outermost transaction in a set of nested transactions is called the top-level
transaction.
 Transactions other than the top-level transaction are called sub-transactions.
 The following are the operations are allowed:
 openSubTransaction(trans) subTrans: Opens a new subtransaction whose
parent is trans and returns a unique subtransaction identifier.
 getStatus(trans)committed, aborted, provisional: Asks the coordinator to
report on the status of the transaction trans. Returns values representing one of
the following: committed, aborted or provisional.
 When a sub-transaction completes, it makes an independent decision either to
commit provisionally or to abort.
 A coordinator for a sub-transaction will provide an operation to open a sub-
transaction, together with an operation enabling that coordinator to enquire
whether its parent has yet committed or aborted.
 Distributed Systems 4.42

Hierarchic Two phase commit protocol

 In this approach, the two-phase commit protocol becomes a multi-level nested
protocol.
 The coordinator of the top-level transaction communicates with the coordinators of
the sub-transactions for which it is the immediate parent.
 It sends canCommit? messages to each of the latter, which in turn pass them on to
the coordinators of their child transactions.
 canCommit?(trans, subTrans) Yes / No: Call from coordinator to coordinator of
child sub-transaction to ask whether it can commit a sub-transaction subTrans. The
first argument, trans, is the transaction identifier of the top-level transaction.
Participant replies with its vote, Yes / No.

Flat Two phase commit protocol

 In this approach, the coordinator of the top-level transaction sends canCommit?
messages to the coordinators of all of the sub-transactions in the provisional
commit list.
 During the commit protocol, the participants refer to the transaction by its top-
level TID.
 Each participant looks in its transaction list for any transaction or sub-transaction
matching that TID.
 A participant can commit descendants of the top-level transaction unless they have
aborted ancestors.
 When a participant receives a canCommit? request, it does the following:
 If the participant has any provisionally committed transactions that are descendants
of the top-level transaction, trans, it:
 checks that they do not have aborted ancestors in the abortList, then
prepares
to commit (by recording the transaction and its objects in permanent
storage);
 aborts those with aborted ancestors;
 sends a Yes vote to the coordinator.
 If the participant does not have a provisionally committed descendent of the top
level transaction, it must have failed since it performed the sub- transaction and it
sends a No vote to the
coordinator
 Distributed Systems 4.43

4.14 DISTRIBUTED DEADLOCKS

• A cycle in the global wait-for graph (but not in any single local one) represents a
distributed deadlock.
• A deadlock that is detected but is not really a deadlock is called a phantom
deadlock.
• Two-phase locking prevents phantom deadlocks; autonomous aborts may cause
phantom deadlocks.
• Permanent blocking of a set of processes that either compete for system resources
or communicate with each other is deadlock.
• No node has complete and up-to-date knowledge of the entire distributed system.
This is the cause of deadlocks.

Fig 4.17: Distributed deadlocks and wait for

graphs Types of distributed deadlock
 Resource deadlock :Set of deadlocked processes, where each process waits for a
resource held by another process (e.g., data object in a database, I/O resource on a
server)
 Communication deadlocks: Set of deadlocked processes, where each process
waits to receive messages (communication) from other processes in the set.
 Distributed Systems 4.44

Local and Global Wait for Graphs

Edge Chasing
 When a server notes that a transaction T starts waiting for another transaction U,
which is waiting to access a data item at another server, it sends a probe containing
TU to the server of the data item at which transaction U is blocked.
 Detection: receive probes and decide whether deadlock has occurred and whether to
forward the probes.
 When a server receives a probe TU and finds the transaction that U is waiting for,
say V, is waiting for another data item elsewhere, a probe TUV is forwarded.
 Detection: receive probes and decide whether deadlock has occurred and whether to
forward the probes.
When a server receives a probe TU and finds the transaction that U is waiting for,
say V, is waiting for another data item elsewhere, a probe TUV is forwarded.
 Resolution: select a transaction in the cycle to abort

Transaction priorities
 Every transaction involved in a deadlock cycle can cause deadlock detection to be
initiated.
 The effect of several transactions in a cycle initiating deadlock detection is that
detection may happen at several different servers in the cycle, with the result that more
than one transaction in the cycle is aborted.

Fig 4.17: Initiated probes

 Distributed Systems 4.45

 Consider transactions T, U, V and W, where U is waiting for Wand V is waiting for T.

 At about the same time, T requests the object held by U and W requests the object held
by V.
 Two separate probes, <T U > and <W V >, are initiated by the servers of these
objects and are circulated until deadlocks are detected by each of the servers.
 The cycle is <T UW V T>and, the cycle is <W V T U W >.
 In order to ensure that only one transaction in a cycle is aborted, transactions are given
priorities in such a way that all transactions are totally ordered.
 In order to ensure that only one transaction in a cycle is aborted, transactions are given
priorities in such a way that all transactions are totally ordered.
 The problem is that the order in which transactions start waiting can determine whether
or not a deadlock will be detected.
 The above pitfall can be avoided by using as scheme in which coordinators save copies
of all the probes received on behalf of each transaction in a probe queue.
 When a transaction starts waiting for an object, it forwards the probes in its queue to
the server of the object, which propagates the probes on down hill routes.

4.14 REPLICATION
Replication in distributed systems enhances the performance, availability and fault
tolerance. The general requirement includes:
• Replication transparency
• Consistency

4.14.1 System Model

Fig 4.18: Architecture for replication management

 The data in a system consist of a collection of items called objects.
 Distributed Systems 4.46

 An object could be a file, say, or a Java object.

 But each such logical object is implemented by a collection of physical copies
called replicas.
 The replicas are physical objects, each stored at a single computer, with data and
behavior that are tied to some degree of consistency by the system‟s operation.
 The system models include replica managers which are components that contain
the replicas on a given computer and perform operations upon them directly.
 Each client‟s requests are first handled by a component called a front end.
 The role of the front end is to communicate by message passing with one or
more of the replica managers, rather than forcing the client to do this itself
explicitly.
 It is the vehicle for making replication transparent.
 A front end may be implemented in the client‟s address space, or it may be a
separate process.
Phases in request processing:
• Issuance of request: The front end issues the request to one or more replica
managers:
– either the front end communicates with a single replica manager, which in turn
communicates with other replica managers
– or the front end multicasts the request to the replica managers.

• Coordination:
 The replica managers coordinate in preparation for executing therequest
consistently.
 They agree, if necessary at this stage, on whether the request is to be applied.
 They also decide on the ordering of this request relative to others.
 The types of ordering includes: FIFO ordering, casual ordering and total
ordering.
• Execution: The replica managers execute the request – perhaps tentatively: that is, in
such a way that they can undo its effects later.
• Agreement: The replica managers reach consensus on the effect of the request – if
any – that will be committed.
• Response: One or more replica managers respond to the front end.
 Distributed Systems 4.47

4.15 CASE STUDY: CODA

Coda is a distributed file system. Coda has been developed at Carnegie Mellon
University (CMU) in the 1990s, and is now integrated with a number of popular UNIX-based
operating systems such as Linux.

Features of Coda File System (CFS):

 CFS main goal is to achieve high availability.
 It has advanced caching schemes.
 It provide transparency
Architecture of Coda
 The clients cache entire files locally.
 Cache coherence is maintained by the use of callbacks. Clients dynamically find
files on server and cache location information.
 For security, token-based authentication and end-to-end encryption is used.

Fig 4.18: Coda file systems

 It is a principle of the design of Coda that the copies of files residing on servers
are more reliable than those residing in the caches of clients.
 It is possible to construct a file system that relies entirely on cached copies of
files in client.
 But such systems will have poor QOS.
 Distributed Systems 4.48

 The Coda servers exist to provide the necessary quality of service.

 The copies of files residing in client caches are regarded as useful only as long as
their currency can be revalidated against the copies in server.
 Revalidation occurs when disconnected operation ceases and the cached files are
reintegrated with those in the servers.
Communication in Coda

Fig 4.19: Communication in Coda

 Coda uses RPC2: a sophisticated reliable RPC system.
 The RPC2 start a new thread for each request, server periodically informs client it is
still working on the request.
 RPC2 is useful for video streaming.
 RPC2 also has multicast support
 A side effect is a mechanism by which the client and server can communicate using
an application-specific protocol.
 Coda uses side effect mechanism.
 Coda servers allow clients to cache whole files.
 Modifications by other clients are notified through invalidation messages require
multicast RPC. The modifications can be done in any one of the following ways:
 Sending an invalidation message one at a time
 Sending invalidation messages in parallel
 Distributed Systems 4.49

Processes in Coda
Cod a maintains distinction between client and server processes. The clients are known
as Venus pr ocesses and Server as Vice processes. The threads are non preemptive and operate
entirely in user space. Low-level thread handles I/O operations.
Caching in Coda
 Cache consistency in Coda is maintained using callbacks.
 The Vice server tracks all clients that have a copy of the file and provide
callback promise. The tokens are obtained from vice server.
 It guarantee that Venus will be notified if file is modified

Upon modification Vice server send invalidate to clients.

Fig 4.20: Caching in Coda

Server Replication in Coda
 The basic unit of replication is termed as volume.
 The Volume Storage Group (VSG) is set of servers that have a copy of a volume.
 The Accessible Volume Storage Group (AVSG) is set of servers in VSG that the client
can contact .
 The Coda uses vector versioning:
 One entry for each server in VSG

 When file updated, corresponding version in AVSG is updated

 Distributed Systems 4.50

Fig 4.21: Server Replication

 In the above figure, the versioning vector at the time of partition is :[1,1,1]
 Client A updates file : versioning vector in its partition: [2,2,1]
 Client B updates file: versioning vector in its partition: [1,1,2]
 After the partition is repaired, compare versioning vectors. There will be a
conflict.

Fault tolerance

Fig 4.22: Fault Tolerance

The following are the fault tolerating actions:
 HOARDING: File cache in advance with all files that will be accessed when
disconnected
 EMULATION: when disconnected, behavior of server emulated at client
 REINTEGRATION: transfer updates to server; resolves conflicts
 Distributed Systems 4.51
REVIEW QUESTIONS
PART – A
1. Define clock
skew.
Clock skew is defined as the difference between the times on two clocks.
2. Define clock drift.
Clock drift is the count time at different rates.
3. What is Clock drift rate ?
Clock drift rate is the difference in precision between a prefect reference clock and a
physical clock.
4. What is External synchronization?
 This method synchronize the process‟s clock with an authoritative external
reference clock S(t) by limiting skew to a delay bound D > 0 - |S(t) - Ci(t) | <
D for all t.
 For example, synchronization with a UTC (Coordinated Universal
Time)source.
5. What is Internal synchronization?
 Synchronize the local clocks within a distributed system to disagree by not
more than a delay bound D > 0, without necessarily achieving external
synchronization - |Ci(t) - Cj(t)| < D for all i, j, t n
 For a system with external synchronization bound of D, the internal
synchronization is bounded by 2D.
6. Give the types of clocks.
Two types of clocks are used:
 Logical clocks : to provide consistent event ordering
 Physical clocks : clocks whose values must not deviate from the real time by
more than a certain amount.
7. What are the techniques are used to synchronize clocks?
 time stamps of real-time clocks
 message passing
 round-trip time (local measurement)
 Distributed Systems 4.52

8. List the algorithms that provides clock synchronization.

 Cristian‟s algorithm
 Berkeley algorithm
 Network time protocol (Internet)
9. Give the working of Berkley’s algorithm.
The time daemon asks all the other machines for their clock values. The machines
answer the request. The time daemon tells everyone how to adjust their clock.
10. What is NTP?
The Network Time Protocol defines architecture for a time service and a protocol to
distribute time information over the Internet.
11. Give the working of Procedure call and symmetric modes:
 All messages carry timing history information.
 The history includes the local timestamps of send and receive of the previous
NTP message and the local timestamp of send of this message
 For each pair i of messages (m, m‟) exchanged between two servers the
following values are being computed
- offsetoi : estimate for the actual offset between two clocks
- delay di : true total transmission time for the pair of messages.
12. Define casual ordering.
The partial ordering obtained by generalizing the relationship between two process is
called as happened-before relation or causal ordering or potential causal ordering.
13. Define logical clock.
A Lamport logical clock is a monotonically increasing software counter, whose
value need bear no particular relationship to any physical clock.
14. What is global state?
The global state of a distributed system consists of the local state of each process,
together with the messages that are currently in transit, that is, that have been sent
but not delivered.
15. When do you call an object to be a garbage?
An object is considered to be garbage if there are no longer any references to it
anywhere in the distributed
system.
 Distributed Systems 4.53

16. Define distributed deadlock.

A distributed deadlock occurs when each of a collection of processes waits for

another process to send it a message, and where there is a cycle in the graph of this
„waits-for‟ relationship.

17. What is distributed snapshot?

Distributed Snapshot represents a state in which the distributed system might have
been in. A snapshot of the system is a single configuration of the system.

18. What is consistent cut?

A consistent global state is one that corresponds to a consistent cut.

19. Define run.

A run is a total ordering of all the events in a global history that is consistent with
each local history‟s ordering.

20. What is linearization?

A linearization or consistent run is an ordering of the events in a global history that

is consistent with this happened-before relation o on H.

21. What is global state predicate?

A global state predicate is a function that maps from the set of global states of
processes in the system.

22. What are the features of the unreliable failure detectors?

• unsuspected or suspected (i.e.) there can be no evidence of failure

• each process sends ``alive'' message to everyone else

• not receiving ``alive'' message after timeout

• This is present in most practical systems

23. What are the features of the reliable failure detectors?

• Unsuspected or failure

• They are present in synchronous system

24. Define distributed mutual exclusion.

Distributed mutual exclusion provide critical region in a distributed environment.

 Distributed Systems 4.54

25. What are the requirements for Mutual Exclusion (ME)?

[ME1] safety: only one process at a
time [ME2] liveness: eventually enter or
exit
[ME3] happened-before ordering: ordering of enter() is the same as HB ordering
26. What are the criteria for performance measures?
Bandwidth consumption, which is proportional to the number of messages sent in
each entry and exit operations.
The client delay incurred by a process at each entry and exit operation.
Throughput of the system: Rate at which the collection of processes as a whole can
access the critical section.
27. What is ring based algorithm?
This provides a simplest way to arrange mutual exclusion between N processes
without requiring an additional process is arrange them in a logical ring.
28. What is mutual synchronisation?
This exploits mutual exclusion between N peer processes based upon multicast.
Processes that require entry to a critical section multicast a request message, and can
enter it only when all the other processes have replied to this message.
29. What is Maekawa’s Voting Algorithm?
In this algorithm, it is not necessary for all of its peers to grant access. Only need to
obtain permission to enter from subsets of their peers, as long as the subsets used by
any two processes overlap.
30. What is election algorithm?
An algorithm for choosing a unique process to play a particular role is called an
election algorithm.
31. What is Ring based Election Algorithm?
 All the processes arranged in a logical ring.
 Each process has a communication channel to the next process.
 All messages are sent clockwise around the ring.
 Assume that no failures occur, and system is asynchronous.
 The ultimate goal is to elect a single process coordinator which has the largest
identifier
 Distributed Systems 4.55

32. What is bully algorithm?

This algorithm allows process to crash during an election, although it assumes the
message delivery between processes is reliable.
33. What are the messages in Bully algorithm?
There are three types of messages:
a. Election message: This is sent to announce an election message. A process
begins an election when it notices, through timeouts, that the coordinator has
failed. T=2Ttrans+Tprocess From the time of sending
b. Answermessage: This is sent in response to an election message.
c. Coordinatormessage: This is sent to announce the identity of the elected
process.
34. Define transaction.
A Transaction defines a sequence of server operations that is guaranteed to be
atomic in the presence of multiple clients and server crash.
35. List the methods to ensure serializability.
There are three ways to ensure serializability:
 Locking
 Timestamp ordering
 Optimistic concurrency control
36. Give the advantages of nested transactions.
 Sub- transactions at same level can run concurrently.
 Sub- transactions can commit or abort independently.
37. Give the rules for committing of nested transactions.
 A transaction may commit or abort only after its child transactions have
completed.
 When a sub-transaction completes, it makes an independent decision either to
commit provisionally or to abort. Its decision to abort is final.
 When a parent aborts, all of its sub-transactions are aborted.
 When a sub -transaction aborts, the parent can decide whether to abort or not.
 If the top-level transaction commits, then all of the sub-transactions that have
provisionally committed can commit too, provided that none of their ancestors
has

aborted.
 Distributed Systems 4.56

38. What are the information held by the locks?

Each instance of Lock maintains the following information in its instance
variables:
 the identifier of the locked object
 the transaction identifiers of the transactions that currently hold the lock
 a lock type
39. What is two phase locking?
The basic two-phase locking (2PL) protocol states:
 A transaction T must hold a lock on an item x in the appropriate mode before T
accesses x.
 If a conflicting lock on x is being held by another transaction, T waits.
 Once T releases a lock, it cannot obtain any other lock subsequently.
40. Define deadlock.
Deadlock is a state in which each member of a group of transactions is waiting for
some other member to release a lock.
41. Give the disadvantages of locking and serialization.
 Lock maintenance represents an overhead that is not present in systems that do
not support concurrent access to shared data. Locking sometimes are only
needed for some cases with low probabilities.
 The use of lock can result in deadlock. Deadlock prevention reduces
concurrency severely. The use of timeout and deadlock detection is not ideal
for interactive programs.
 To avoid cascading aborts, locks cannot be released until the end of the
transaction. This may reduce the potential for concurrency.
42. What is multi version time stamp ordering?
In multi-version timestamp ordering, a list of old committed versions as well as
tentative versions is kept for each object. This list represents the history of the values
of the object. The benefit of using g multiple versions is that read operations that
arrive too late need not be rejected.
43. What is Hierarchic Two phase commit protocol?
 In this approach, the two-phase commit protocol becomes a multi-level nested
protocol.
 The coordinator of the top-level transaction communicates with the
coordinators of the sub-transactions for which it is the immediate parent.
 Distributed Systems 4.57

44. What is Flat Two phase commit protocol?

 In this approach, the coordinator of the top-level transaction sends canCommit?
messages to the coordinators of all of the sub-transactions in the provisional
commit list.
 During the commit protocol, the participants refer to the transaction by its top-
level TID.
 Each participant looks in its transaction list for any transaction or sub-
transaction matching that TID.

45. Define phantom deadlock.

A deadlock that is detected but is not really a deadlock is called a phantom deadlock.

46. What is replication?

Replication in distributed systems enhances the performance, availability and fault
tolerance.

47. What is Coda?

Coda is a distributed file system. Coda has been developed at Carnegie Mellon
University (CMU) in the 1990s, and is now integrated with a number of popular
UNIX-based operating systems such as Linux.

48. What are the fault tolerating actions?

 HOARDING: File cache in advance with all files that will be accessed when
disconnected
 EMULATION: when disconnected, behaviour of server emulated at client
 REINTEGRATION: transfer updates to server; resolves conflicts

PART - B
1. Explain the clocking in detail.
2. Describe Cristian‟s algorithm.
3. Write about Berkeley algorithm
4. Brief about NTP.
5. Explain about logical time and logical clocks.
6. Describe global states.
 Distributed Systems 4.58

7. Brief about distributed mutual exclusion.

8. Write in detail about election algorithms.
9. Explain transactions and concurrency control.
10. Describe nested transaction and its issues.

11. What is optimistic concurrency control?

12. Write in detail about timestamp ordering.

13. Describe atomic commit protocol.

14. Explain distributed deadlocks.
15. Describe replication.
16. Write about Coda.
 PROCESS AND RESOURCE
5
MANAGEMENT
5.1 PROCESS MANAGEMENT

Process Management involves the activities aimed at defining a

process, establishing responsibilities, evaluating process performance,
and identifying opportunities for improvement.

5.1.1 Process Migration

• Process migration involves the transfer of sufficient amount of the state of a

process from one computer to another.

Process migration is the act of transferring process between two machines.

• The process executes on the target machine.

• The aim for migration is to move processes from heavily loaded to lightly load
systems.
• This enhances the performance communication.
• Processes that interact intensively can be moved to the same node to reduce
communications cost.
• The hope of process migration is that it will allow for better system-wide
utilization of resources.
• The long-running process may be moved because the machine it is running on will
be down.
• Process can take advantage of unique hardware or software capabilities.
5.2 Process and Resource Management
• Process migration enables dynamic load distribution, fault resilience, eased system
administration, and data access locality.
• After migration, the process on the source system and create it on the target
system.
• The process image and process control block and any links must be moved.

Machine S Machine D

1
d
a
2
c 2
b
1 c
b c
a 4 d
f e
5a a 3
b
4a
P1 P2 P3 P4 P5 P1 P2 P3

Kernel Kernel

Fig 5.1 Before Migration

Machine S Machine D

1
d
a2

c 2
b
1 c b c
a 4 d
f e

5a a 3

b a
4
P1 P2 P1 P2 P3 P4

Kernel Kernel

Fig 5.2: After Migration

Issues in migration:
The following questions must be addressed in process migration:
– Who initiates the migration?

– What is involved in a Migration?

– What portion of the process is migrated?

– What happens to outstanding messages and signals?

 Distributed 5.3
System
Moving PCBs:
 The movement of the process control block is straightforward and simple.
 The difficulty, from a performance point of view, concerns the process address space
and any open files assigned to the process.

 There are several strategies for moving the address space and data including:

 Eager (All):
 This transfers the entire address space.
 No trace of process is left behind in the original system.
 If address space is large and if the process does not need most of it,
then this approach is unnecessarily expensive.
 Implementations that provide a checkpoint/restart facility are likely to
use this approach, because it is simpler to do the check-pointing and
restarting if all of the address space is localized.
 Precopy
 In this method, the process continues to execute on the source node
while the address space is copied to the target node.
 The pages modified on the source during the precopy operation have to
be copied a second time.
 This strategy reduces the time that a process is frozen and cannot
execute during migration.
 Eager (dirty)
 The transfer involves only the portion of the address space that is in
main memory and has been modified.
 Any additional blocks of the virtual address space are transferred on
demand.
 The source machine is involved throughout the life of the process.
 Copy-on-reference
 In this method, the pages are only brought over when referenced.
 This has the lowest initial cost of process migration.
5.4 Process and Resource Management Distributed 5.4
System
 Flushing
 The pages are cleared from main memory by flushing dirty pages to
disk.
 This relieves the source of holding any pages of the migrated process in
main memory.
Negotiation in Migration
• The selection of migration policy is the responsibility of Starter utility.
• Starter utility is also responsible for long-term scheduling and memory allocation.
• The decision to migrate must be reached jointly by two Starter processes (i.e.) the
process on the source and on the destination machines.

Fig 5.3: Negotiation in Migration process

Eviction
 Sometimes the destination system may refuse to accept the migration of a process
to itself.
 If a workstation is idle, process may have been migrated to it.
 Once the workstation is active, it may be necessary to evict (recover) the migrated
processes to provide adequate response time.
 Distributed 5.5
System
Distributed Global States
• The operating system cannot know the current state of all process in the
distributed system.
• A process can only know the current state of all processes on the local system.
• Also the remote processes only know state information that is received by
messages.
• This message helps to know about the state in the past.

5.2 THREADS

A thread is the entity within a process that can be scheduled for execution. All
threads of a process share its virtual address space and system resources.

Each thread maintains exception handlers, a scheduling priority, thread local storage, a
unique thread identifier, and a set of structures the system will use to save the thread context
until it is scheduled. The thread context includes the thread's set of machine registers, the
kernel stack, a thread environment block, and a user stack in the address space of the thread's
process. Threads can also have their own security context, which can be used for
impersonating clients.
Because threads have some of the properties of processes, they are sometimes called
lightweight processes. In a process, threads allow multiple executions of streams. A process
can be single threaded or multi- threaded.

Fig 5.4: Threads

Threads provide better responsiveness, resource sharing, economy and better utilization
of multiprocessor architectures.
 Distributed 5.6
(2) Create new thread System
to
(1) request service the request

Client Server Thread

(3) Resume listening for

additional client request

Fig 5.5: Multithreaded Architectures

Thread management done by user-level threads library. The actions in thread
management involve:
 Creation and destruction of threads
 Message passing or data sharing between threads
 Scheduling threads
 saving/restoring threads contexts
5.2.1 Multithreading Models
There are basically three multithreading models based on how the user and kernel
threads map into each other:
 Many to one
 One to one
 Many to many

Many to one Model

Fig 5.6: Many to one Model

 Distributed 5.7
System
In this model many user-level threads mapped to single kernel thread. Examples are
Solaris Green Threads and GNU Portable Threads.

One to One Model

K K K K

Fig 5.7: One to One Model

In this model each user-level thread maps to kernel thread. The disadvantage of this
model is creating a user thread requires creating a kernel thread. Example are threads in
Windows NT/XP/2000, Linux, Solaris 9 and later.

Many to many model

K K K

Fig 5.8: Many to many model

This model allows many user level threads to be mapped to many kernel threads. This
also allows the operating system to create a sufficient number of kernel threads. Examples
includes Solaris prior to version 9 and Windows NT/2000 with the ThreadFiber package.

Two Level Model

This is similar to Many to Many model, except that it allows a user thread to be bound
to a kernel thread. Examples are threads in IRIX, HP-UX, Tru64 UNIX, Solaris 8 and earlier.
 Distributed 5.8
System

K K K K
Fig 5.9: Two Level Model
5.2.2 Threading Issues
The following are the issues in threads:
 Semantics of fork() and exec() system calls
The fork() call may duplicate either the calling thread or all
threads.
 Thread cancellation
 Terminating a thread before it has finished.
 There are two general approaches:
 Asynchronous cancellation terminates the target thread immediately
 Deferred cancellation allows the target thread to periodically check if it
should be cancelled
 Signal handling
 Signals are used in UNIX systems to notify a process that a particular event
has occurred
 A signal handler is used to process signals
 Signal is generated by particular event
 Signal is delivered to a process
 Signal is handled
 The signal can be delivered :
 to the thread to which the signal applies
 Distributed 5.9
System
 to every thread in the process

 to certain threads in the process

 Assign a specific thread to receive all signals for the process.
 Thread pools
 Create a number of threads in a pool where they await work.
 This is slightly faster to service a request with an existing thread than create a
new thread
 This allows the number of threads in the application(s) to be bound to the size
of the pool.
 Thread specific data

 This allows each thread to have its own copy of data

 This is useful when you do not have control over the thread creation process
(i.e., when using a thread pool)
 Scheduler activations
 The Many to Many models require communication to maintain the appropriate
number of kernel threads allocated to the application.
 The scheduler activations provide upcalls: a communication mechanism from
the kernel to the thread library.
 This communication allows an application to maintain the correct number
kernel threads.
5.2.3 Thread Implementation
The various thread implementations are listed below:

 PThreads
 This is a POSIX standard (IEEE 1003.1c).
 This is a API for thread creation and synchronization and specifies the
behavior of the thread library.
 pthreads is an Object Orientated API that allows user-land multi-threading in
PHP.
 It includes all the tools needed to create multi-threaded applications targeted at
the Web or the Console.
 Distributed 5.1
System 0
 PHP applications can create, read, write, execute and synchronize with
Threads, Workers and Threaded objects.
 A Threaded Object:
 A Threaded Object forms the basis of the functionality that allows
pthreads to operate.
 It exposes synchronization methods and some useful interfaces for the
programmer.
 Thread:
 The user can implement a Thread by extending the Thread declaration
provided by pthreads implementing the run method.
 Any members of the thread can be written and read by any context
with a reference to the Thread.
 Any context can be executed in any public and protected methods.
 The run method of the implementation is executed in a separate thread
when the start method of the implementation is called from the context
that created it.
 Only the context that creates a thread can start and join with it.
 Worker Object:
 A Worker Thread has a persistent state, and will be invoked by start().
 The life of worker thread is till the calling object goes out of scope, or
is explicitly shutdown.
 Any context with a reference can stack objects onto the Worker, which
will be executed by the Worker in a separate Thread.
 The run method of a Worker is executed before any objects on the
stack, such that it can initialize resources that the objects to come may
need.
 Pool:
 A Pool of Worker threads can be used to distribute Threaded objects
among Workers.
 The Pool class included implements this functionality and takes care of
referencing in a sane manner.
 The Pool implementation is the easiest and most efficient way of using
multiple threads.
 Distributed 5.1
System 1
 Synchronization:
 All of the objects that pthreads creates have built in synchronization in
the (familiar to java programmers ) form of ::wait and ::notify.
 Calling ::wait on an object will cause the context to wait for another
context to call ::notify on the same object.
 This allows for powerful synchronization between Threaded Objects in
PHP.
 Method Modifiers:
 The protected methods of Threaded Objects are protected by pthreads,
such that only one context may call that method at a time.
 The private methods of Threaded Objects can only be called from
within the Threaded Object during execution.
 Data Storage:
 Any data type that can be serialized can be used as a member of a
Threaded object, it can be read and written from any context with a
reference to the Threaded Object.
 Not every type of data is stored serially, basic types are stored in their
true form. Complex types, Arrays, and Objects that are not Threaded
are stored serially; they can be read and written to the Threaded Object
from any context with a reference.
 With the exception of Threaded Objects any reference used to set a
member of a Threaded Object is separated from the reference in the
Threaded Object; the same data can be read directly from the Threaded
Object at any time by any context with a reference to the Threaded
Object.
 Static Members:
 When a new context is created , they are generally copied, but
resources and objects with internal state are nullified .
 This allows them to function as a kind of thread local storage.
 Allowing the new context to initiate a connection in the same way as
the context that created it, storing the connection in the same place
without affecting the original context.
 These threads are common in UNIX operating systems (Solaris, Linux,
Mac OS
X).
 Distributed 5.1
System 2
 Windows Threads

 Microsoft Windows supports preemptive multitasking, which creates the effect

of simultaneous execution of multiple threads from multiple processes.

 On a multiprocessor computer, the system can simultaneously execute as many

threads as there are processors on the computer.

 A job object allows groups of processes to be managed as a unit.

 Job objects are namable, securable, sharable objects that control attributes of
the processes associated with them.

 Operations performed on the job object affect all processes associated with the
job object.
 An application can use the thread pool to reduce the number of application
threads and provide management of the worker threads.
 Applications can queue work items, associate work with waitable handles,
automatically queue based on a timer, and bind with I/O.

 User-mode scheduling (UMS) is a lightweight mechanism that applications

can use to schedule their own threads.

 An application can switch between UMS threads in user mode without

involving the system scheduler and regain control of the processor if a UMS
thread blocks in the kernel.

 Each UMS thread has its own thread context instead of sharing the thread
context of a single thread.

 The ability to switch between threads in user mode makes UMS more efficient
than thread pools for short-duration work items that require few system calls.
 A fiber is a unit of execution that must be manually scheduled by the
application.

 Fibers run in the context of the threads that schedule them.

 Each thread can schedule multiple fibers.

 In general, fibers do not provide advantages over a well-designed

multithreaded application.

 However, using fibers can make it easier to port applications that were
designed to schedule their own
threads.
 Distributed 5.1
System 3
 Java Threads
 Java is a multi threaded programming language which means we can develop
multi threaded program using Java.
 A multi threaded program contains two or more parts that can run concurrently
and each part can handle different task at the same time making optimal use of
the available resources especially when your computer has multiple CPUs.
 Java threads may be created by: Extending Thread class or by implementing the
Runnable interface.
 Java threads are managed by the JVM.
Runnable run()

New Thread () New Running

Dead

Waiting
Fig 5. 10: Lifecycle of Java Thread
 Thread class provide constructors and methods to create and perform operations
on a thread.
 Thread class extends Object class and implements Runnable interface.
 The Runnable interface should be implemented by any class whose instances
are intended to be executed by a thread.
 Runnable interface have only one method named run().

5.3 RESOURCE MANAGEMENT IN DISTRIBUTED SYSTEMS

 Distributed systems contain a set of resources interconnected by a network.
 Processes are migrated to fulfill their resource requirements.
 The Resource manager is responsible to control the assignment of resources to
processes.
 The resources can be logical (shared file) or physical (CPU) resource.
 Scheduling is the way in which processors are assigned to run on the available
resources.
 Distributed 5.1
System 4
 In a distributed computing system, the scheduling of various modules on particular
processing nodes may be preceded by appropriate allocation of modules of the
different tasks to various processing nodes and then only the appropriate execution
characteristic can be obtained.
 The task allocation becomes the important most and major activity in the task
scheduling within the operating system of a DCS.
5.3.1 Features of Scheduling algorithms
The following are the desired features of scheduling algorithms:
 General purpose
 A scheduling approach should make few assumptions about and have few
restrictions to the types of applications that can be executed.
 Interactive jobs, distributed and parallel applications, as wellas non-
interactive batch jobs, should all be supported with good performance.
 This property is a straightforward one, but to some extent difficult to
achieve.
 Because different kinds of jobs have different attributes, their requirements
to the scheduler may contradict.
 To achieve the general purpose, a tradeoff may have to be made.
 Efficiency:
 It has two meanings: one is that it should improve the performance of
scheduled jobs as much as possible; the other is that the scheduling
should incur reasonably low overhead so that it would not counter
attack the benefits.
 Fairness:
 Sharing resources among users raises new challenges in guaranteeing
that each user obtains his/her fair share when demand is heavy is
fairness.
 In a distributed system, this problem could be exacerbated suchthat one
user consumes the entire system.
 There are many mature strategies to achieve fairness on a single node.
 Dynamic:
 The algorithms employed to decide where to process a task should
respond to load changes, and exploit the full extent of the resources
available.
 Distributed 5.1
System 5
 Transparency:
 The behavior and result of a task’s execution should not be affected by
the host(s) on which it executes.
 In particular, there should be no difference between local and remote
execution.
 No user effort should be required in deciding where to execute a task
or in initiating remote execution; a user should not even be aware of
remote processing.
 Further, the applications should not be changed greatly.
 It is undesirable to have to modify the application programs in order to
execute them in the system.
 Scalability
 A scheduling algorithm should scale well as the number of nodes
increases.
 An algorithm that makes scheduling decisions by first inquiring the
workload from all the nodes and then selecting the most lightly loaded
node has poor scalability.

 This will work fine only when there are few nodes in the system.
 This is because the inquirer receives a flood of replies almost
simultaneously, and the time required to process the reply messages for
making a node selection is too long as the number of nodes (N)
increase.
 Also the network traffic quickly consumes network bandwidth.
 A simple approach is to probe only m of N nodes for selecting a node.

 Fault tolerance
 A good scheduling algorithm should not be disabled by the crash of one
or more nodes of the system.
 Also, if the nodes are partitioned into two or more groups due to link
failures, the algorithm should be capable of functioning properly for the
nodes within a group.
 Algorithms that have decentralized decision making capability and
consider only available nodes in their decision making have better fault
tolerance
capability.
 Distributed 5.1
System 6
 Quick decision making capability
 Heuristic methods requiring less computational efforts (and hence less
time) while providing near-optimal results are preferable to exhaustive
(optimal) solution methods.
 Balanced system performance and scheduling overhead
 Algorithms that provide near-optimal system performance with a
minimum of global state information (such as CPU load) gathering
overhead are desirable.
 This is because the overhead increases as the amount of global state
information collected increases.
 This is because the usefulness of that information is decreased due to
both the aging of the information being gathered and the low
scheduling frequency as a result of the cost of gathering and processing
the extra information.
 Stability
 Fruitless migration of processes, known as processor thrashing, must be
prevented.
 E.g. if nodes n1 and n2 observe that node n3 is idle and then offload a
portion of their work to n3 without being aware of the offloading
decision made by the other node.
 Now if n3 becomes overloaded due to this it may again start
transferring its processes to other nodes.
 This is caused by scheduling decisions being made at each node
independently of decisions made by other nodes.
5.3.2 Task Assignment Approach
The following assumptions are made in this approach:
 A process has already been split up into pieces called tasks.
 This split occurs along natural boundaries (such as a method), so that each task
will have integrity in itself and data transfers among the tasks are minimized.
 The amount of computation required by each task and the speed of each CPU
are known. The cost of processing each task on every node is known. This is
derived from assumption 2.
 The IPC cost between every pair of tasks is already known.
 The IPC cost is 0 for tasks assigned to the same node.
 Distributed 5.1
System 7
 This is usually estimated by an analysis of the static program.
 If two tasks communicate n times and the average time for each inter-task
communication is t, them IPC costs for the two tasks is n * t.
 Precedence relationships among the tasks are known.
 Reassignment of tasks is not possible.
Goal of this method is to assign the tasks of a process to the nodes of a distributed
system in such a manner as to achieve goals such as:
 Minimization of IPC costs
 Quick turnaround time for the complete process
 A high degree of parallelism
 Efficient utilization of system resources in general
• These goals often conflict. E.g., while minimizing IPC costs tends to assign all
tasks of a process to a single node, efficient utilization of system resources tries
to distribute the tasks evenly among the nodes.
• So also, quick turnaround time and a high degree of parallelism encourage
parallel execution of the tasks, the precedence relationship among the tasks
limits their parallel execution.
• In case of m tasks and q nodes, there are mq possible assignments of tasks to
nodes .
• In practice, however, the actual number of possible assignments of tasks to
nodes may be less than mq due to the restriction that certain tasks cannot be
assigned to certain nodes due to their specific requirements.
5.3.3 Classification of Load Balancing Algorithms

Fig 5.11: Classification of Load Balancing Algorithms

 Distributed 5.1
System 8
 Static versus Dynamic

Static Algorithms Dynamic Algorithms

Static algorithms use only information Dynamic algorithms collect state
about the average behavior of the system. information and react to system state if it
changed.
Static algorithms ignore the current state Dynamic algorithms are able to give
or load of the nodes in the system. significantly better performance.
Static algorithms are much simpler. They are complex

 Deterministic versus Probabilistic

Deterministic Algorithms Probabilistic Algorithms

Deterministic algorithms use the Probabilistic algorithms use information
information about the properties of the of static attributes of the system (e.g.
nodes and the characteristic of processes number of nodes, processing capability,
to be scheduled. topology) to formulate simple process
placement rules
Deterministic approach is difficult to Probabilistic approach has poor
optimize. performance

 Centralized versus Distributed

Centralized Algorithms Distributed Algorithms

Centralized approach collects information Distributed approach contains entities to
to server node and makes assignment make decisions on a predefined set of
decision. nodes
Centralized algorithms can make efficient Distributed algorithms avoid the
decisions, have lower fault-tolerance bottleneck of collecting state information
and react faster

 Cooperative versus Non-cooperative

Cooperative Algorithms Non-cooperative Algorithms

In Co-operative algorithms distributed In Non-cooperative algorithms entities
entities cooperatewith each other. act as autonomous ones and make
scheduling decisions independently from
other entities.
 Distributed 5.1
System 9
Cooperative algorithms are more They are simpler.
complex and involve larger overhead.

Stability of Cooperative algorithms are Stability is comparatively poor.

better.

Issues in Load Balancing Algorithms

 Load estimation policy: determines how to estimate the workload of a node.
 Process transfer policy: determines whether to execute a process locally or
remote.
 State information exchange policy: determines how to exchange
load information among nodes.
 Location policy: determines to which node the transferable process should be
sent.
 Priority assignment policy: determines the priority of execution of local and
remote processes.
 Migration limiting policy: determines the total number of times a process can
migrate.
Load Estimation
Policies: Policy I:
 To balance the workload on all the nodes of the system, it is necessary to
decide how to measure the workload of a particular node.
 Some measurable parameters (with time and node dependent factor) can be the
following:
 Total number of processes on the node
 Resource demands of these processes
 Instruction mixes of these processes
 Architecture and speed of the node’s processor
 Several load-balancing algorithms use the total number of processes to achieve big
efficiency.
 Distributed 5.2
System 0
Policy II:
 In some cases the true load could vary widely depending on the remaining service time,
which can be measured in several way:
 Memory less method assumes that all processes have the same
expected remaining service time, independent of the time used so far.
 Past repeats assumes that the remaining service time is equal to the
time used so far.
 Distribution method states that if the distribution service times is
known, the associated process’s remaining service time is the expected
remaining time conditioned by the time already used.
Policy III:
 None of the previous methods can be used in modern systems because of
periodically running processes and daemons.
 An acceptable method for use as the load estimation policy in these systems would
be to measure the CPU utilization of the nodes.
 Central Processing Unit utilization is defined as the number of CPU cycles actually
executed per unit of real time.
 It can be measured by setting up a timer to periodically check the CPU state
(idle/busy).
Threshold Policy
 Most of the algorithms use the threshold policy to decide on whether the node is
lightly-loaded or heavily-loaded.
 Threshold value is a limiting value of the workload of node which can be
determined by:
 Static policy: predefined threshold value for each node depending on processing
capability.
 Dynamic policy: threshold value is calculated from average workload and a
predefined constant
 Below threshold value node accepts processes to execute, above threshold value
node tries to transfer processes to a lightly-loaded node.
Single threshold Policy
 Single-threshold policy may lead to unstable algorithm because under loaded node
could turn to be overloaded right after a process migration.
 Distribute d 5.21
System
 To reduce instability double-threshold policy has been propos ed which is also
known as high-low policy.

Fig 5.12: Process Threshold Policy

Double threshold Policy
 When node is in overloaded region new local processes are sent to run remotely,
requests to accept remote processes are rejected.
 When node is in normal region new local processes run locally, requests to accept
remote processes are rejected.
 When node is in under loaded region new local processes run locally, requests to
accept remote processes are accepted
Location Policies:
 Threshold method:
This policy selects a random node, checks whether the node is able to receive the
process, then transfers the process. If node rejects, another node is selected
randomly. This continues until probe limit is reached.

 Shortest method
 L distinct nodes are chosen at random, each is polled to determine its load. The
process is transferred to the node having the minimum value unless its
workload value prohibits to accept the process.
 Simple improvement is to discontinue probing whenever a node with zero load
is encountered.
 Bidding method
 The nodes contain managers (to send processes) and contractors (to receive
processes).
 Distributed 5.22
System
 The managers broadcast a request for bid, contractors respond with bids (prices
based on capacity of the contractor node) and manager selects the best offer.
 Winning contractor is notified and asked whether it accepts the process for
execution or not.
 The full autonomy for the nodes regarding scheduling.
 This causes big communication overhead.
 It is difficult to decide a good pricing policy.
 Pairing
 Contrary to the former methods the pairing policy is to reduce the variance of
load only between pairs.
 Each node asks some randomly chosen node to form a pair with it.
 If it receives a rejection it randomly selects another node and tries to pair again.
 Two nodes that differ greatly in load are temporarily paired with each other and
migration starts.
 The pair is broken as soon as the migration is over.
 A node only tries to find a partner if it has at least two processes.
Priority assignment policy
 Selfish: Local processes are given higher priority than remote processes. Worst
response time performance of the three policies.
 Altruistic: Remote processes are given higher priority than local processes. Best
response time performance of the three policies.
 Intermediate: When the number of local processes is greater or equal to the
number of remote processes, local processes are given higher priority than remote
processes. Otherwise, remote processes are given higher priority than local
processes.
Migration limiting policy
This policy determines the total number of times a process can migrate
 Uncontrolled: A remote process arriving at a node is treated just as a process
originating at a node, so a process may be migrated any number of times
 Controlled:
 Avoids the instability of the uncontrolled policy.
 Distributed 5.23
System
 Use a migration count parameter to fix a limit on the number of time a process
can migrate.
 Irrevocable migration policy: migration count is fixed to 1.
 For long execution processes migration count must be greater than 1 to adapt
for dynamically changing states
5.3.4 Load-sharing approach
 The following problems in load balancing approach led to load sharing approach:
 Load balancing technique with attempting equalizing the workload on all the
nodes is not an appropriate object since big overhead is generated by gathering
exact state information.
 Load balancing is not achievable since number of processes in a node is always
fluctuating and temporal unbalance among the nodes exists every moment
Basic idea:
 It is necessary and sufficient to prevent nodes from being idle while some other
nodes have more than two processes.
 Load-sharing is much simpler than load-balancing since it only attempts to ensure
that no node is idle when heavily node exists.
 Priority assignment policy and migration limiting policy are the same as that for
the load-balancing algorithms.

Load Estimation Policies

 Since load-sharing algorithms simply attempt to avoid idle nodes, it is sufficient to
know whether a node is busy or idle.
 Thus these algorithms normally employ the simplest load estimation policy of
counting the total number of processes.
 In modern systems where permanent existence of several processes on an idle
node is possible, algorithms measure CPU utilization to estimate the load of a
node
Process Transfer Policies
 The load sharing algorithms normally use all-or-nothing strategy.
 This strategy uses the threshold value of all the nodes fixed to 1.
 Nodes become receiver node when it has no process, and become sender node
when it has more than 1 process.
 Distributed 5.24
System
 To avoid processing power on nodes having zero process load-sharing algorithms
uses a threshold value of 2 instead of 1.
 When CPU utilization is used as the load estimation policy, the double-threshold
policy should be used as the process transfer policy

Location Policies
 The location policy decides whether the sender node or the receiver node of the
process takes the initiative to search for suitable node in the system, and this policy
can one of the following:
 Sender-initiated location policy: Sender node decides where to send the process.
Heavily loaded nodes search for lightly loaded nodes
 Receiver-initiated location policy: Receiver node decides from where to get the
process. Lightly loaded nodes search for heavily loaded nodes

Sender-initiated location policy

 Node becomes overloaded, it either broadcasts or randomly probes the other nodes
one by one to find a node that is able to receive remote processes.
 When broadcasting, suitable node is known as soon as reply arrives

Receiver-initiated location policy

 Nodes becomes underloaded, it either broadcast or randomly probes the other
nodes one by one to indicate its willingness to receive remote processes.
 Receiver-initiated policy require preemptive process migration facility since
scheduling decisions are usually made at process departure epochs
 Both policies gives substantial performance advantages over the situation in
which no load-sharing is attempted.
 Sender-initiated policy is preferable at light to moderate system loads.
 Receiver-initiated policy is preferable at high system loads.
 Sender-initiated policy provide better performance for the case when process
transfer cost significantly more at receiver-initiated than at sender-initiated
policy due to the pre-emptive transfer of processes.

State information exchange policies

 In load-sharing algorithms it is not necessary for the nodes to periodically
exchange state information, but needs to know the state of other nodes when it is
either underloaded or
overloaded.
 Distributed 5.25
System
 The following are the two approaches followed when there is state change:

 Broadcast when state changes

 In sender-initiated/receiver-initiated location policy a node broadcasts State
Information Request when it becomes overloaded/ underloaded.
 It is called broadcast-when-idle policy when receiver-initiated policy is used
with fixed threshold value of 1
 Poll when state changes
 In large networks polling mechanism is used.
 Polling mechanism randomly asks different nodes for state information until
find an appropriate one or probe limit is reached.
 It is called poll-when-idle policy when receiver-initiated policy is used with
fixed threshold value value of 1.
 Distributed 5.26
System
REVIEW QUESTIONS
PART - A
1. Define process management.
Process Management involves the activities aimed at defining a process, establishing
responsibilities, evaluating process performance, identifying and opportunities
or improvement.
2. Define process migration.
Process migration is the act of transferring process between two machines.
3. What is Eager (All)?
 This transfers the entire address space.
 No trace of process is left behind in the original system.
 If address space is large and if the process does not need most of it, then this
approach is unnecessarily expensive.
 Implementations that provide a checkpoint/restart facility are likely to use this
approach, because it is simpler to do the check-pointing and restarting if all of the
address space is localized.
4. What is Precopy?
 In this method, the process continues to execute on the source node while the
address space is copied to the target node.
 The pages modified on the source during the precopy operation have to be copied a
second time.
 This strategy reduces the time that a process is frozen and cannot execute during
migration.
5. What is Eager (dirty)?
 The transfer involves only the portion of the address space that is in main memory
and has been modified.
 Any additional blocks of the virtual address space are transferred on demand.
 The source machine is involved throughout the life of the process.
6. What is Copy-on-reference?
 In this method, the pages are only brought over when referenced.
 This has the lowest initial cost of process migration.
 Distributed 5.27
System
7. What is Flushing?
 The pages are cleared from main memory by flushing dirty pages to disk.
 This relieves the source of holding any pages of the migrated process in main
memory.
8. Define thread.
A thread is the entity within a process that can be scheduled for execution. All threads
of a process share its virtual address space and system resources.
9. What is thread context?
The thread context includes the thread's set of machine registers, the kernel stack, a
thread environment block, and a user stack in the address space of the thread's process.
10. What is thread pool?
Thread pool consists of a number of threads in a pool where they await work. This is
slightly faster to service a request with an existing thread than create a new thread.
This allows the number of threads in the application(s) to be bound to the size of the
pool.
11. What is worker thread?
A Worker Thread has a persistent state, and will be invoked by start(). The life of
worker thread is till the calling object goes out of scope, or is explicitly shutdown
12. What is job object?
A job object allows groups of processes to be managed as a unit.
13. What is UMS?
User-mode scheduling (UMS) is a lightweight mechanism that applications can use to
schedule their own threads.
14. What is a fiber?
A fiber is a unit of execution that must be manually scheduled by the application.
15. What is multithreaded program?
A multi- threaded program contains two or more parts that can run concurrently and
each part can handle different task at the same time making optimal use of the
available resources especially when your computer has multiple CPUs.
16. List the features of scheduling algorithm.
 General purpose
 Efficiency
 Distributed 5.28
System
 Fairness
 Dynamic
 Transparency
 Scalability
 Fault tolerance
 Quick decision making capability
 Balanced system performance and scheduling overhead
 Stability
17. Classify load balancing algorithms.

18. Differentiate static and dynamic algorithms.

Static Algorithms Dynamic Algorithms

Static algorithms use only information Dynamic algorithms collect state
about the average behavior of the system. information and react to system state if it
changed.

Static algorithms ignore the current state Dynamic algorithms are able to give
or load of the nodes in the system. significantly better performance.
Static algorithms are much simpler. They are complex
 Distributed 5.29
System
19. Differentiate deterministic and probabilistic algorithms.

Deterministic Algorithms Probabilistic Algorithms

Deterministic algorithms use the Probabilistic algorithms use information
information about the properties of the of static attributes of the system (e.g.
nodes and the characteristic of processes number of nodes, processing capability,
to be scheduled. topology) to formulate simple process
placement rules

Deterministic approach is difficult to Probabilistic approach has poor

optimize. performance

20. Differentiate centralized and distributed algorithms.

Centralized Algorithms Distributed Algorithms

Centralized approach collects information Distributed approach contains entities to
to server node and makes assignment make decisions on a predefined set of
decision. nodes
Centralized algorithms can make efficient Distributed algorithms avoid the
decisions, have lower fault-tolerance bottleneck of collecting state information
and react faster

21. Differentiate cooperative and non co operative algorithms.

Cooperative Algorithms Non-cooperative Algorithms

In Co-operative algorithms distributed In Non-cooperative algorithms entities
entities cooperatewith each other. act as autonomous ones and make
scheduling decisions independently from
other entities.
Cooperative algorithms are more They are simpler.
complex and involve larger overhead.

Stability of Cooperative algorithms are Stability is comparatively poor.

better.
 Distributed 5.30
System
22. Give the issues in Load Balancing Algorithms
 Load estimation policy: determines how to estimate the workload of a node.
 Process transfer policy: determines whether to execute a process locally or remote.
 State information exchange policy: determines how to exchange load information
among nodes.
 Location policy: determines to which node the transferable process should be sent.
 Priority assignment policy: determines the priority of execution of local and
remote processes.
 Migration limiting policy: determines the total number of times a process can
migrate.
23. What is threshold policy?
Most of the algorithms use the threshold policy to decide on whether the node is
lightly-loaded or heavily-loaded. Threshold value is a limiting value of the workload
of node which can be determined by:
 Static policy: predefined threshold value for each node depending on processing
capability.
 Dynamic policy: threshold value is calculated from average workload and a
predefined constant
24. What is single threshold policy?
Single-threshold policy may lead to unstable algorithm because under loaded node
could turn to be overloaded right after a process migration.
25. What is double threshold policy?
When node is in overloaded region new local processes are sent to run remotely,
requests to accept remote processes are rejected. When node is in normal region new
local processes run locally, requests to accept remote processes are rejected.When
node is in under loaded region new local processes run locally, requests to accept
remote processes are accepted
26. What are the location policies?
 Threshold method
 Shortest method
 Bidding method
 Pairing
 Threshold method:
 Distributed 5.31
System
27. What is threshold policy?
This policy selects a random node, checks whether the node is able to receive the
process, then transfers the process. If node rejects, another node is selected randomly.
This continues until probe limit is reached.

28. What is Shortest method?

 L distinct nodes are chosen at random, each is polled to determine its load. The
process is transferred to the node having the minimum value unless its
workload value prohibits to accept the process.
 Simple improvement is to discontinue probing whenever a node with zero load
is encountered.
29. What is Bidding method?
 The nodes contain managers (to send processes) and contractors (to receive
processes).
 The managers broadcast a request for bid, contractors respond with bids (prices
based on capacity of the contractor node) and manager selects the best offer.
 Winning contractor is notified and asked whether it accepts the process for
execution or not.
 The full autonomy for the nodes regarding scheduling.
 This causes big communication overhead.
 It is difficult to decide a good pricing policy.
30. What is Pairing?
 Contrary to the former methods the pairing policy is to reduce the variance of
load only between pairs.
 Each node asks some randomly chosen node to form a pair with it.
 If it receives a rejection it randomly selects another node and tries to pair again.
 Two nodes that differ greatly in load are temporarily paired with each other and
migration starts.
 The pair is broken as soon as the migration is over.
 A node only tries to find a partner if it has at least two processes.
 Distributed 5.32
System
31. What is Priority assignment policy?
 Selfish: Local processes are given higher priority than remote processes. Worst
response time performance of the three policies.
 Altruistic: Remote processes are given higher priority than local processes. Best
response time performance of the three policies.
 Intermediate: When the number of local processes is greater or equal to the number
of remote processes, local processes are given higher priority than remote
processes. Otherwise, remote processes are given higher priority than local
processes.

32. What are the issues in load balancing approach?

The following problems in load balancing approach led to load sharing approach:
 Load balancing technique with attempting equalizing the workload on all the
nodes is not an appropriate object since big overhead is generated by gathering
exact state information.
 Load balancing is not achievable since number of processes in a node is always
fluctuating and temporal unbalance among the nodes exists every moment
33. What is load sharing approach?
It is necessary and sufficient to prevent nodes from being idle while some other nodes
have more than two processes. Load-sharing is much simpler than load-balancing
since it only attempts to ensure that no node is idle when heavily node exists. Priority
assignment policy and migration limiting policy are the same as that for the load-
balancing algorithms.

34. What are location policies?

 The location policy decides whether the sender node or the receiver node of the
process takes the initiative to search for suitable node in the system, and this policy
can one of the following:
 Sender-initiated location policy: Sender node decides where to send the process.
Heavily loaded nodes search for lightly loaded nodes
 Receiver-initiated location policy: Receiver node decides from where to get the
process. Lightly loaded nodes search for heavily loaded nodes
 Distributed 5.33
System
PART-B
1. Explain in detail about process migration.
2. Define threads. Explain all the multithreading thread models.
3. What are the issues in threads?
4. Discuss the various thread implementations.
5. Describe the resource management in distributed systems.
6. Write about task assignment.
7. Classify load balancing algorithms.
8. Explain load sharing approaches.