Introduction to Distributed Systems

Introduction to Distributed Systems

 Key concepts

 What is Distributed System?

 Fundamental issues in Distributed Systems

 Distributed System models and Architectures

 Design goals of Distributed Systems

 Differences between client/server and D.S. environments

2/36
What is distributed system?

 Consist of several computers that do not share a memory

or a clock. {Excludes tightly coupled Systems.}
 Communicate with each other by message passing over a
communication network.
 Each computer has its own memory & run its own
operating systems.
 The design of Distributed Operating System is much more
difficult due to the lack of both shared memory and a
physical clock, and unpredictable communication delays.

3/36
 What is distributed system?
 Definition
 Distributed System consists of a collection of autonomous
computers linked by a computer network and equipped with
distributed system software.
 Distributed System software enables computers to co-ordinate
their activities & to share the resources of the system –
hardware, software and data.
 Users of a Distributed System should perceive a single,
integrated computing facility even through it may be
implemented by many computers in different locations.
 Entails the dispersion of hardware, software & data to
multiple computers

4/36
What is distributed system?
 DS definitions
 A collection of autonomous computers linked by a
computer network and supported by software that
enables the collections to operate as an integrated
facility.
 Example: ATM, WWW

 What is distributed?
 Processing logic

 Functions

 Data

 Control

5/36
Characteristics of Distributed Systems
 Multiple nodes working independently with each others.
{NODE → CPU & MEMORY}

 The nodes do not have common shared memory.

 There exists an interconnection network

 Distribution transparency is provided

 There is no single point of failure. { fault tolerance}

6/36
 Advantages of distributed systems over
centralized systems
• Economics: a collection of microprocessors offer a better
price/performance than mainframes. Low price/performance
ratio: cost effective way to increase computing power.

• Speed: a distributed system may have more total computing

power than a mainframe. Ex. 10,000 CPU chips, each running
at 50 MIPS. Not possible to build 500,000 MIPS single
processor since it would require 0.002 nsec instruction cycle.
Enhanced performance through load distributing.

• Inherent distribution: Some applications are inherently

distributed. Ex. a supermarket chain.

7/36
 Advantages of distributed systems over
centralized systems
• Reliability: If one machine crashes, the system as a whole can
still survive. Higher availability and improved reliability.

• Incremental growth: Computing power can be added in small

increments. Modular expandability

• Another deriving force: the existence of large number of

personal computers, the need for people to collaborate and
share information.

8/36
Advantages of distributed systems over
independent PCs
 Data sharing:
 allow many users to access to a common data base
 Resource Sharing:
 expensive peripherals like color printers
 Communication:
 enhance human-to-human communication
 e.g., email, chat
 Flexibility:
 spread the workload over the available machines

9/36
Disadvantages of Distributed Systems
 Software:
 difficult to develop software for distributed systems
 Network:
 saturation, lossy transmissions
 Security:
 easy access also applies to secrete data

10/36
Significant Consequences of Distributed Systems
 Concurrency
 The capacity of the system to handle shared resources can be
increased by adding more resources (for example Computers) to
the network.
 No global clock
 There is no single global notion of the correct time. The only
communication is by sending messages through a network.
 Independent failures
 In DS, faults in the network result in the isolation of the
computers that are connected to it.
 The programs may not be able to detect whether the network has
failed or has become unusually slow.
11/36
Distributed Systems Challenges

Heterogeneity

Openness

Security

Scalability

Failure handling
Concurrency

Transparency

12/36
Heterogeneity

 Different networks, hardware, operating systems,

programming languages, developers.
 Need to set up protocols to solve these heterogeneities.
 Middleware: a software layer that provides a
programming abstraction as well as masking the
heterogeneity.
 Mobile code: code that can be sent from one computer to
another and run at the destination.

13/36
Openness
 The openness of DS is determined primarily by the degree
to which new resource-sharing services can be added and
be made available for use by a variety of client programs.
 Open systems are characterized by the fact that their key
interfaces are published.
 Open DS are based on the provision of a uniform
communication mechanism and published interfaces for
access to shared resources.
 Open DS can be constrcted from heterogeneous hardware
and software.

14/36
Security
 Security for information resources has three
components:
 Confidentiality: protection against disclosure to
unauthorized individuals.
 Integrity: protection against alteration or corruption.
 Availability: protection against interference with the
means to access the resources.
 Two new security challenges:
 Denial of service attacks (DoS).
 Security of mobile code.

15/36
Scalability
 A system is described as scalable if it remains effective when
there is a significant increase in the number of resources and
the number of users.
 Challenges:
 Controlling the cost of resources.
 Controlling the performance loss.
 Preventing software resources from running out
 Avoiding preformance bottlenecks.

16/36
Failure handling
 When faults occur in hardware or software, programs may
produce incorrect results or they may stop before they have
completed the intended computation.
 Techniques for dealing with failures:
 Detecting failures
 Masking failures
 Tolerating failures
 Recovering form failures
 Redundancy

17/36
Concurrency
 There is a possibility that several clients will attempt to
access a shared resource at the same time.
 Any object that represents a shared resource in a distributed
system must be responsible for ensuring that it operates
correctly in a concurrent environment.

18/36
 Transparency
 Transparency is defined as the concealment from the user and
the application programmer of the separation of components
in a distributed system, so that the system is perceived as a
whole rather than as a collection of independent components.
 Eight forms of transparency:
 Access transparency
 Location transparency
 Concurrency transparency
 Replication transparency
 Failure transparency
 Mobility transparency
 Performance transparency
 Scaling transparency
19/36
Issues in designing distributed O.S.
 Transparency
 Kinds of transparency
 Access Transparency: local and remote resources are

accessed in the same way.

 Location Transparency: The location of objects is not known

to the users.
 Name transparency
 User mobility

 Replication (&Fragmentation) Transparency: enables

multiple instances of resources to be used to increase
reliability and performance without the knowledge of the
replicas by the user or application programmer. Objects are
accessed without knowledge about any possible
fragmentations.
20/36
Issues in designing distributed O.S. (contd…)
 Concurrency transparency: Objects are accessed in the same
way as in a single – user system.

 Error (Failure) Transparency: Applications terminate in a

defined way even in the event of hardware or software failures.
 Migration(mobility) transparency: Objects can migrate without
affecting applications.
 Performance transparency: the system automatically
reconfigures to improve performance.
 Scaling transparency: The System can be extended in size
without changes to its structures or the applications.

21/36
Issues in designing distributed O.S. (contd…)
 Reliability
 Methods for dealing with faults, and to detect and remove
the faults:
 Fault avoidance
 Fault tolerance
 Redundancy techniques
 Distributed control
 Fault detection and recovery
 Atomic transactions
 Stateless servers
 Acknowledgments
22/36
Issues in designing distributed O.S. (contd…)
 Flexibility
 Design of a distributed operating system should be flexible
due to following reasons
 Ease of modification
 Ease of enhancement
 Performance
 Design principles
 Batch if possible
 Cache whenever possible

23/36
Issues in designing distributed O.S. (contd…)
 Minimize copying of data
 Minimize network traffic
 Take advantage of fine-grain parallelism for multiprocessing
 Scalability
 Avoid centralized entities
 Avoid centralized algorithms
 Perform most operations on client workstations
 Heterogeneity
 Security
 Emulation of existing OS

24/36
Distributed system models and architectures
 Models are classified into five categories

1. MINICOMPUTER MODEL

2. WORKSTATION MODEL

3. WORKSTATION-SERVER MODEL

4. PROCESSOR-POOL MODEL

5. HYBRID MODEL

25/36
Minicomputer model
 Simple extension of centralized time sharing system
 May be used when resource sharing with remote users is
desired
 Consists of few minicomputers interconnected
 Each minicomputer has multiple users simultaneously logged
on to it through several interactive terminals connected to the
minicomputer
 Each user logged on to one specific minicomputer, with
remote access to other minicomputers and remote shared
resources
 Example: ARPAnet
26/36
Minicomputer model
Mini
Computer
TERMINALS

Communication
Mini Mini
Network
Computer Computer

Mini
Computer

27/36
Workstation model
 Several workstations are interconnected by a
communication network

 User logs onto one of the workstations (“home”) and

submits jobs for execution.

 Each workstation is serving as a single user computer

 If workstation does not have sufficient power for

execution

 It transfers one or more processes from user’s

workstation to some other workstation that is idle
and gets work done

28/36
Workstation model
Workstation

Workstation
Workstation

Communication
Network

Workstation
Workstation

Workstation

29/36
Workstation model( Contd…)

 Issues in Workstation model are

 How does the system find an idle workstation?
 How is a process transferred from one workstation to get it
executed on another workstation?
 What happens to a remote process if a user logs on a
workstation that has idle until now and was being used to
execute a process of another workstation?

30/36
Workstation-Server model

 Is a network of several personal workstations, each

with its own disk and local file system and few
minicomputers
 User logs onto a workstation called his/her home
workstation
 Normal computing activities required by the user’s
processes are performed at users’s workstation
 Requests for services provided by special servers are
sent to servers providing that service

31/36
Workstation-server model
Workstation Workstation
(diskless) (diskfull)

Workstation
Workstation (diskless)
(diskless)

Communication
Network
Minicomputer as
Mini Computer as Data base Server
file server

Minicomputer as Minicomputer as
Print Server Email Server

32/36
Workstation-Server vs Workstation model
 It is much cheaper – use of few minicomputers with large fast
disks than large number of diskful workstations with small,
slow disk
 Backup and hardware maintenance are easier to perform
with a few large disks than many small disks scattered
around
 Since all files are managed by the file servers, users have the
flexibility to use any workstation and access the files
 Process migration facility is not needed here
 Model does not utilize the processing capability of idle
workstations
 A user has guaranteed response time because workstations
are not used for executing remote processes
33/36
Processor-pool model
 Processors are pooled together to be shared by the users
as needed

 Each processor in the pool has its own memory to load

and run a system program or an application program of
distributed computing system

 No concept of Home computer

 Compared to workstation server model, this model allows

better utilization of the available processing power of ta
distributed computing system

34/36
Processor-pool model
Terminal
Terminal Terminal

Terminal
Terminal
Communication Network

Terminal

Run File
Ser Ser
Pool of
ver … … ver
processors

35/36
Hybrid Model
 Combines the advantages of both workstation-server and
processor pool model

 Efficient execution of computation-intensive jobs

 Gives guaranteed response to interactive jobs by allowing

them to be processed locally

 More expensive to build

36/36
Why are distributed systems gaining popularity?
 Inherently distributed applications

 Information sharing among distributed users

 Resource sharing

 Better price-performance ratio

 Shorter response times and higher throughput

 Higher reliability

 Extensibility and incremental growth

 Better flexibility in meeting user’s needs

37/36
End of chapter 1

38/36