Chapter 1

Introduction

1
 1. Introduction
1.1 Introduction to Distributed Systems
1.2 Characteristics of Distributed Systems
1.3 Goals of Distributed Systems
1.4 Hardware Concepts
1.5 Software Concepts
1.6 Multiprocessor Systems, Multicomputer
Systems
1.7 Distributed Programming

2
 Computer and Network Evolution
Computer Systems
– 10 million dollars and 1 instruction/sec

– 1000 dollars and 1 billion instructions/sec

Computer Networks
– Local-area networks (LANs)
• Small amount of information, a few microseconds

• Large amount of information, at rate of 100 million to 10 billion bits/sec

– Wide-area networks (WANs)
• 64 Kbps gigabits per second

3
Definition of a Distributed System

A distributed system is:

A collection of independent computers

that appears to its users as a single
coherent system

4
Definition of a Distributed System

1.1

A distributed system organized as middleware

Note that the middleware layer extends over multiple machines5
Characteristics of Distributed Systems
• Fault-Tolerant: Distributed systems consist of a large number
of hardware and software modules that are bound to fail in the
long run. Such component failures can lead to service
unavailability.

• Scalable: A distributed system can operate correctly even as

some aspect of the system is scaled to a larger size.

• Security: Distributed systems should allow communication

between programs/users/ resources on different computers by
enforcing necessary security arrangements.

• Transparency: Distributed systems should be perceived by

users and application developers as a whole rather than as a 6
collection of cooperating components.
 1.2 Goals of Distributed Systems

 Making resources available

 Distribution transparency
 Openness
 Scalability

7
 Make Resources Accessible
 Access resources and share them in a controlled
and efficient way.
 Printers, computers, storage facilities, data, files, Web
pages, and networks, …
 Connecting users and resources also makes it
easier to collaborate and exchange information.
 Internet for exchanging files, mail, documents, audio,
and video

8
 Openness

Goal: Open distributed system -- able to interact with

services from other open systems, irrespective of the
underlying environment:
 System should conform to well-defined interfaces

 Systems should support portability of applications

 Systems should easily interoperate

 Achieving openness: At least make the distributed system

independent from heterogeneity of the underlying

environment:
 Hardware

 Platforms

 Languages

9
 Scale in Distributed Systems
 Scalability: At least three components:
 Number of users and/or processes (size scalability)
 Maximum distance between nodes (geographical scalability)
 Number of administrative domains (administrative scalability)

10
 Techniques for Scaling
 Distribution: Partition data and
computations across multiple machines:
 Move computations to clients (Java applets)
 Decentralized naming services (DNS)
 Decentralized information systems (WWW)

11
Scaling Techniques (cont..)

1.4

The difference between letting:

a) a server or
b) a client check forms as they are being filled
12
13
 Techniques for Scaling
 Replication/caching: Make copies of data
available at different machines:
 Replicated file servers and databases
 Mirrored Web sites

14
 Transparency in a
Distributed System
Transparency Description

Hide differences in data representation and how a resource is

Access
accessed
Location Hide where a resource is located
Migration Hide that a resource may move to another location
Hide that a resource may be moved to another location while
Relocation
in use

Replication Hide that a resource may be replicated on multiple servers

Hide that a resource may be shared by several competitive

Concurrency
users
Failure Hide the failure and recovery of a resource

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

15
Advantages of distributed systems
over centralized systems

16
Advantages of distributed systems
over Independent PCs

17
 Disadvantages of distributed
systems
Security and privacy
•Security: The prevention and protection of (computer) assets
from unauthorized access.
•Privacy: The right of the individual to be protected against
intrusion into personal affair by direct physical means or by
publication of information.
“The most secure computers are those not connected to the
Internet and shielded from any interference”

18
 Developing Distributed Systems: Pitfalls

 There are many false assumptions:

 The network is reliable
 The network is secure
 The network is homogeneous
 The topology does not change
 Latency is zero
 Bandwidth is infinite
 Transport cost is zero
 There is one administrator

19
 Hardware and Software concepts

 Distributed system may be

 Multi-processor System (tightly coupled system)
 Multi-computers System (loosely coupled system)

 Major issues
 How they are interconnected?
 How they communicate?

20
 Hardware Concepts

 Tightly Coupled versus Loosely Coupled

 A tightly-coupled system is one where the components tend to be
reliably connected in close proximity.
 A loosely-coupled system is one where the components tend to be
distributed.
 Tightly coupled systems (multiprocessors)
o shared memory

o intermachine delay short, data rate high

 Loosely coupled systems (multicomputers)

o private memory

o intermachine delay long, data rate low

21
 Models of Computing

Memory
(Code and Data)

Instructions Data

I/O with external

world
CPU

The Von Neumann architecture

22
Shared Memory in Multiprocessors

Shared Memory
(Code and Data)

I D I D I D I D

CPU CPU CPU CPU

Parallel Random Access Machine model:

•N processors connected to shared memory
•All memory addresses reachable in unit time by
any CPU

23
 Shared Memory & Private Memory
 Multiprocessors (not multicomputers)
 Single physical address space shared by all CPUs
 CPU A writes 37 to address 1000

 CPU B then reads from address 1000 and gets 37

 e.g., multiple processors on a board with shared

memory
 Multicomputers
 Every machine has its own private memory
 CPU A writes 37 to its address 1000

 CPU B reads from its address 1000 and gets whatever

happens to be there; not affected by the other write

 e.g., PCs connected by a network

24
 Bus-based & Switch-based

 Bus architecture of the interconnection network

 single network, backplane, bus, cable or other medium
that connects all the machines
 e.g., cable television
 Switched architecture
 Msgs move along wires with an explicit switching
decision made at each step to route the message
along one of the outgoing wires.
 e.g., worldwide public telephone system

25
 Multicomputers

 Bus-Based multicomputers
 easy to build

 communication volume much smaller

 Switched multicomputers
 interconnection networks

26
• A bus can get overloaded rather quickly with each CPU accessing
the bus for all data and instructions.
• A solution to this is to add cache memory between the
CPU and the bus.
• The cache holds the most recently accessed regions of memory.
• This way, the CPU only has to go out to the bus to access main
memory only when the regions are not in its cache.

28
 Switched Multiprocessors
 Using switches enables us to achieve a far greater CPU
density in multiprocessor systems.
 An m ×n crossbar switch is a switch that allows any of m
elements to be switched to any of n elements.
 To use a crossbar switch, we place the CPUs on one axis
(e.g. m) and the break the memory into a number of chunks
which are placed on the second axis (e.g. n memory
chunks).
 There will be a delay only when multiple CPUs try to access
the same memory chunk.

29
Switched Multiprocessors

30
 Software Concept
 The software design goal in building a distributed system is to
create a Single System Image - have a collection of
independent computers appear as a single system to the user(s).

 By single system, we refer to creating a system in which

the user is not aware of the presence of multiple
computers or of distribution.
 Two types
 Loosely coupled software
 Tightly-coupled software

31
 Software Concept
 Loosely-coupled - software in which the systems interact with
each other to a limited extent as needed. For the most part, they
operate as fully-functioning stand-alone machines.
 If the network goes down, things are pretty much functional.
 Loosely coupled systems may be ones in which there are shared
services (parts of file service, web service).
 With tightly-coupled software, there is a strong dependence on
other machines for all aspects of the system.
 Essentially, both the interconnect and functioning of the remote
systems are necessary for the local system's operation.

32
 Software-Hardware Combination

 Three types:
 Network Operating Systems

 (True) Distributed Systems

 Multiprocessor Time Sharing

33
 Network Operating Systems

 loosely-coupled software on loosely-coupled

hardware
 Eg. A network of workstations connected by LAN
 Each machine has a high degree of autonomy
 Files servers: client and server model

34
 (True) Distributed Systems
 tightly-coupled software on loosely-coupled hardware
 provide a single-system image or a virtual machine
(uniprocessor)
 To accomplish this, we need certain capabilities:
 Uniform naming from anywhere; the file system
should look the same.
 Same system call interface everywhere

35
 Time Sharing Systems

 Multiple jobs are executed by switching the CPU

between them.
 In this, the CPU time is shared by different
processes, so it is called as “Time sharing
Systems”.
 Time slice is defined by the OS, for sharing CPU
time between processes.

36
 Comparison between Systems

Distributed OS

Item Multiproc Network OS

Multicomp
(Time
(true)
Sharing)
Degree of transparency Very High High Low

Same OS on all nodes Yes Yes No

Number of copies of OS 1 N N
Global,
Resource management Global, central Per node
distributed
Scalability No Moderately Yes

 A comparison between multiprocessor operating systems,

multicomputer operating systems, and network operating systems

37