Chapter 1

Introduction
1.1 Introduction and Definition
 before the mid-80s, computers were
 very expensive (hundred of thousands or even
millions of dollars)
 very slow (a few thousand instructions per second)
 not connected among themselves
 after the mid-80s: two major developments
 cheap and powerful microprocessor-based
computers appeared
 computer networks
 LANs at speeds ranging from 10 to 1000 Mbps
 WANs at speed ranging from 64 Kbps to
gigabits/sec
 consequence
 feasibility of using a large network of computers to
work for the same application; this is in contrast to
08.05.25  2
the old centralized systems where there was a single
What is a distributed system ?
 A distributed system is:
 A collection of autonomous/independent
computing elements that appears to its
users as a single coherent system.

 This definition refers two features of

distributed system
 Computing elements of distributed system are
independent
 Users or applications perceive a single system

3
 Other Definitions

A distributed system is a system designed to support the

development of applications and services which can exploit

a physical architecture consisting of multiple, autonomous

processing elements that do not share primary memory but

cooperate by sending asynchronous messages over a

communication network (Blair & Stefani)

08.05.25
  4
Collection of autonomous nodes
 Autonomous computing elements also called
nodes
 Can be of either HW or SW
 Behave independently of each other
 Each node have its own notion of time
 There is no global clock.
 Leads to fundamental synchronization and coordination
problems.
 However, these nodes collaborate to achieve
a common goal
 Realized by exchanging messages with each
other

5
Coherent system
 A single coherent system – users believe they
are dealing with a single system
 The difference between components as well as the
communication between them are hidden from users
 Users can interact in a uniform and consistent way
regardless of where and when interaction takes place

 Examples
 An end user cannot tell where a computation is taking place
 Where data is exactly stored should be irrelevant to an
application
 If or not data has been replicated is completely hidden
 Note:- Autonomous elements have to
collaborate to achieve this goal

6
 Examples of distributed system
• Peer to Peer System

08.05.25  7
• Datacenter

08.05.25  8
 What is centralized systems
Centralized computing is computing done at a central location,
using terminals that are attached to a central computer.
The computer itself may control all the peripherals directly (if they
are physically connected to the central computer), or they may be
attached via a terminal server

08.05.25  9
 Centralized vs Distributed Systems
Centralized
One component with non-autonomous parts
 Component shared by users all the time
 All resources accessible
 Software runs in a single process
 Single point of control
 Single point of failure
Distributed
Multiple autonomous components
 Components are not shared by all users
 Resources may not be accessible
 Software runs in concurrent processes on different processors
 Multiple points of control and
 Multiple points of failure
08.05.25  10
 Parallel Systems
• Parallel Computing is a form of computation in which many
computations are carried out simultaneously, operating on the
principle that large problems can often be divided into smaller
ones, which are then solved concurrently (in parallel).
• In parallel computing, all processors have access to a shared
memory. Shared memory can be used to exchange information
between processors, tightly coupled systems.
• In distributed computing, each processor has its own private
memory (distributed memory). Information is exchanged by
passing messages between the processors, loosely coupled
systems.

08.05.25
  11
 Why Distributed?
 Resource and Data Sharing
 printers, databases, multimedia servers, ...
 Availability, Reliability
 the loss of some instances can be hidden
 Scalability, Extensibility
 the system grows with demand (e.g., extra servers)
 Performance
 huge power (CPU, memory, ...) available
 Inherent distribution, communication
 organizational distribution, e-mail, video

08.05.25
  12
 Problems of Distribution
 Concurrency, Security
 clients must not disturb each other
 Privacy
 e.g., when building a preference profile
 unwanted communication such as spam
 Partial failure
 we often do not know where the error is (e.g., RPC)
 Location, Migration, Replication
 clients must be able to find their servers
 Heterogeneity
 hardware, platforms, languages, management

08.05.25
  13
 Characteristics of Distributed Systems
 differences between the computers and the ways they
communicate are hidden from users
 users and applications can interact with a distributed
system in a consistent and uniform way regardless of
location
 distributed systems should be easy to expand and scale
 a distributed system is normally continuously available,
even if there may be partial failures
• Certain common characteristics can be used to assess distributed systems.
 Heterogeneity
 Openness
 Security
 Scalability
 Failure Handling
 Concurrency
08.05.25 Transparency  14
  to support heterogeneous computers and networks and to
provide a single-system view, a distributed system is often
organized by means of a layer of software called middleware
that extends over multiple machines.

a distributed system organized as middleware; note that the

middleware layer extends over multiple machines, and offers each
application the same interface
08.05.25
  15
Goals of a Distributed System
 Should you build distributed systems just because you
can?
 Should you build distributed systems for problems that
can be solved by a single machine?
 NO
 There are 4 goals that should be met to make
building a distributed system worth the effort.
 Resource Accessibility - Easy to access and share
resource
 Distribution Transparency - Hide the fact that resources
are distributed across the network
 Openness - The system should offer services according to
standard rules that describe their syntax and semantics
 Extensible: easy to add / replace components
 Scalability - Size scalable, geographically scalable,
administratively scalable
16
Resource sharing/accessibility
 An important goal of distributed system is to make it
easy for users(and applications) to access and share
remote resources.
 Resources can be virtually anything,
 Typical examples, storage facilities, data, files, services, and
networks
 Benefits
 Economic
 It is cheaper to have a single high-end reliable storage facility be
shared than having to buy and maintain storage for each user
separately.
 Encourage collaboration and exchange of information
 allowed geographically dispersed people work together by means
of groupware software such as
 Collaborative editing, teleconferencing, and so on
 Bit Torrent – allows users to share files across the Internet

 Challenges of resource sharing is Security,

 E.g. email spam, DOS attacks
17
Distribution Transparency
 DS hide the fact that its processes and
resources are physically distributed across
multiple computers
 DS that is able to present itself to users and
applications as if it were only a single
computer system is said to be transparent
 That is, invisible, to end users and applications.
 The concept of transparency can be applied
to several aspects of a distributed system.
 The most important ones are shown on the next
slide.

18
 Types of Transparency in a Distributed System
Transparenc Description
y
Access Hide differences in data representation and how an object is
accessed
Location Hide where an object is located
Relocation Hide that an object may be moved to another location while
in use
Migration Hide that an object may move to another location
Replication Hide that an object is replicated
Concurrency Hide that an object may be shared by several independent
users simultaneously
Failure Hide the failure and recovery of an object
Figure 1-2. Different forms of transparency in a distributed system (ISO, 1995) .
The term object refers to either process or resource
19
Questions

 To hide replication from user is it necessary that all

replicas have the same name?

 Would you think that it is possible to implement

replication transparency without location
transparency ?

20
Openness
 An open distributed system is essentially a system
that offers components that can easily be used by, or
integrated into other systems.
 It often consists of component that originate from somewhere
as well
 How?
 Components adhere to standard rules that describe the
syntax and semantics of what services those components
have to offer
 Define services through Interfaces using interface definition
language(IDL)
 It captures only the syntax of services like function name,
parameter, return type etc
 Semantics (what those services do) specified in natural
language

21
Openness
 By openness we want to achieve
 Interoperability
 Implementations from different manufacturers can work
together by merely relying on the standard rules
 Portability
 Applications from one distributed system can be executed
on another distributed system that implements the same
interface
 Extensibility
 Easy to add or replaces components in the system
 Flexibility
 Easy to modify/customize the system/component to a
specific need
 Flexibility is achieved by separating policy from
mechanism
22
Scalability
 Scalability is the ability of a system, network, or process
to handle a growing amount of work in a capable
manner(with no significant loss of performance)
 Measured in three dimensions
 Size scalable
 Can easily add more users or resources to the system
 Geographically scalable
 Can easily handle users and resources that may lie far
apart
 Administratively scalable
 Can easily managed even if it spans many independent
administrative organizations

23
Size Scalability problems
 When services are implemented/running by means of a
single/few tightly coupled servers in the distributed
system.
 The server, or group of servers, can simply become a
bottleneck when it needs to process an increasing number of
requests.
 Root causes for size scalability problems with centralized
solutions
 The computational capacity, limited by the CPUs
 The storage capacity, including the transfer rate between
CPUs and disks(I/O transfer rate)
 The network between the user and the centralized service

24
Size Scalability problems - Examples

Figure 1-3. Examples of scalability limitations.

25
Problems with geographical scalability
 It is difficult to scale existing distributed systems that
were designed for local-area networks into WAN
 Many distributed systems assume synchronous
client-server interactions:
 Client sends request and waits for an answer.
 Latency may easily prohibit this scheme.
 WAN links are often inherently unreliable:
 The effect is that solutions developed for local-area
networks cannot always work on wide-area system
 Example, simply moving streaming video from LAN to WAN
is bound to fail.
 Lack of multipoint communication(like broadcast),
 Unlike LAN, a simple search broadcast cannot be deployed.
 Solution is to develop separate naming and directory
services to find hosts and resources

26
Scaling techniques
 There are basically three techniques for scaling:
1) Hiding communication latencies,
2) Partitioning and Distribution
3) Replication

27
Scaling techniques
1) Hiding communication latencies, it is
applicable in the case of geographical
scalability.
 The basic idea is try to avoid waiting for
responses to remote service request as much
as possible.
 Make use of asynchronous communication
 Make request and do other useful work up until the
response turned in
 Problem: not every application fits this model.
 Reduce the overall communication between client and
server
 Move computations to client
 See next slide

28
Scaling Techniques (1)- Hide latency

Figure 1-4. The difference between letting (a) a server or (b) a

client check forms as they are being filled.
29
Scaling Techniques
2) Partitioning and Distribution
 Involves taking a component, splitting it into
smaller parts, and subsequently spreading those
parts across the system.
 Example: Domain Name System (DNS),
 See next slide

30
Scaling Techniques (2)- Distribution

Figure 1-5. An example of dividing the DNS

31 name space into zones.
Scaling techniques
3) Redundancy/Replication
 Is to make copies of data available at different machines
 Replication has the following advantages
 Hide much of the communication latency problems mentioned
 Improves availability
 Helps to balance the load between components leading to better
performance
Caching is a special form of replication
-is the process of storing copies of files in a cache, or temporary
storage location, so that they can be accessed more quickly
Caching is a decision made by the client of a resource and
not by the owner of a resource.
Drawback
Caching and replication leads to consistency problems

32
Pitfalls when Developing Distributed Systems
 False assumptions made by first time developer:
• 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.
• Note:-
• When one of these fail, it is difficult to mask unwanted
behavior
• Most of these issues will not most likely show up in non-
distributed applications development

33
Types of distributed systems
 Three types of distributed systems
1. High performance distributed computing systems
2. Distributed information systems
3. Distributed systems for pervasive computing

34
1. Distributed Computing Systems
 Used for high performance computing tasks
 Examples distributed computing systems
 Cluster computing systems
 Grid computing systems
 Cloud computing systems

35
Cluster Computing Systems
 In virtually all cases, cluster computing is used for parallel
programming in which a single (compute intensive) program
is run in parallel on multiple machines.
 A typical cluster system consists of a collection of compute
nodes that are controlled and accessed by means of a single
master node.
 Nodes are connected through a LAN.
 Nodes are essentially homogeneous (Same OS, near-identical
hardware )
 The master node typically handles
 The allocation of nodes to a particular parallel program,
 Maintains a batch queue of submitted jobs, and
 Provides an interface for the users of the system.

 Batch processing is a technique for automating and processing multiple transactions as a single group.

36
Example cluster configuration
General configuration of Linux-based cluster called
Beowulf

Figure. An example of a cluster computing system.

•Note:- there is also another configuration (symmetric approach) of Cluster system
where there is no master node. A good example is MOSIX
37
Grid Computing Systems
 In contrast to cluster computing, grid computing
 Have a high degree of heterogeneity
 Dispersed across several organizations
 Can easily span a wide-area network
 The goal is to bring users and resources from different
organizations together to allow collaboration among
them
 Establish Virtual Organization (VO)
 Members belong to the same VO has access to the resources
that are provided to that VO
 Focus of the software design for grid computing
 Providing access to resources from different administrative
domain to only those users that belong to a specific VO
 Typically, resources are
 Compute servers (including supercomputers, cluster computers),
 Storage facilities and Databases.
 Networked telescopes, sensors, etc

38
Cloud computing
 Provides computing services and resources
(hardware and software) over a network/internet
 Cloud computing is based upon the concept of utility
computing
 Customers shall pay only based on a pay-per-use
model
 It is characterized by an easily usable and
accessible pool of virtualized resources
 It is scalable as users can get more resources if more
work needs to be done

39
Cloud computing
 In practice, clouds are organized into four layers
 Hardware – contains resources customers never get to see
directly.
 Infrastructure - Employs virtualization techniques to provide
customers virtual storage and computing resources
 Platform – provides API (Application Programming Interface )for developing
Apps, Storage
 Application – Actual applications like suite of apps shipped
with OSes

40
2. Distributed Information Systems
 Enterprises might already have multiple networked
information systems
 One that makes its services available to remote clients.
 Example a University might have Registrar system, HRM
system etc….
 However, integrating applications into enterprise
wide information system was painful
 Middleware is the solution
 Applications can be integrated at different levels
 Database level
 That results in Distributed Transaction Processing
 Application level
 Enterprise Application Integration (EAI)
41
 Transaction Processing Systems
 Consider database applications
 special primitives are required to program
transactions, supplied either by the underlying
distributed system or by the language runtime system
 exact list of primitives depends on the type of
application
Primitive Description
BEGIN_TRANSACTION Mark the start of a transaction
Terminate the transaction and try to
END_TRANSACTION
commit
ABORT_TRANSACTIO Kill the transaction and restore the
N old values
Read data from a file, a table, or
READ
otherwise
Write data to a file, a table, or
WRITE
08.05.25 otherwise 42

 The Transaction Model
 the model for transactions comes from the world of
business
 a supplier and a retailer negotiate on
 price
 delivery date
 quality
 etc.
 until the deal is concluded they can continue
negotiating or one of them can terminate
 but once they have reached an agreement they are
bound by law to carry out their part of the deal
 transactions between processes is similar with this
scenario
08.05.25  43
  e.g. reserving a seat from DD to Bhardar through AA and
Mekelle airports

BEGIN_TRANSACTION BEGIN_TRANSACTION
reserve DD  AA; reserve DD  AA;
reserve AA  mekelle; reserve AA  mekelle;
reserve mekelle  bhardar; reserve mekelle  bhardar full 
END_TRANSACTION ABORT_TRANSACTION
(a) (b)

(a) transaction to reserve three flights commits

(b) transaction aborts when third flight is unavailable

08.05.25
  44
 properties of transactions(ACID)
1. Atomic: to the outside world, the transaction happens
indivisibly; a transaction either happens completely or not at all;
intermediate states are not seen by other processes
2. Consistent:
- transaction does not violate system invariants;
- ensures that corruption or errors in your data do not create
unintended consequences for the integrity of your table.
3. Isolated or Serializable:
- when multiple users are reading and writing from the same
table all at once, isolation of their transactions ensures that the
concurrent transactions don't affect one another.
4. Durable:
- ensures that changes to your data made by successfully
executed transactions will be saved, even in the event
of system failure; see later in Chapter 8

08.05.25
  45
 Classification of Transactions
 a transaction could be flat, nested or distributed
 Flat Transaction
 consists of a series of operations that satisfy the ACID
properties
 simple and widely used but with some limitations
 do not allow partial results to be committed or
aborted
 atomicity is also partly a weakness
 in our airline reservation example, we may want
to accept the first two reservations and find an
alternative one for the last
 some transactions may take too much time

08.05.25
  46
  Nested Transaction
 constructed from a number of sub transactions; it is
logically decomposed into a hierarchy of sub
transactions
 the top-level transaction forks off children that run in
parallel, on different machines; to gain performance or
for programming simplicity
 each may also execute one or more sub transactions
 permanence (durability) applies only to the top-level
transaction; commits by children should be undone
 Distributed Transaction
 a flat transaction that operates on data that are
distributed across multiple machines
 problem: separate algorithms are needed to handle
the locking of data and committing the entire
08.05.25
 transaction; 47

 (a) a nested transaction
(b) distributed transaction

08.05.25
  48
  Enterprise Application Integration
 how to integrate applications independent from their
databases
 transaction systems rely on request/reply
 how can applications communicate with each other

middleware as a communication facilitator in enterprise application

08.05.25

integration 49
  there are different communication models
 RPC (Remote procedure Call)
 RMI (Remote Method Invocation)
 MOM (Message-Oriented Communication)
 see later in Chapter 4
3. Distributed Pervasive Systems
 the distributed systems discussed so far are
characterized by their stability; fixed nodes having
high-quality connection to a network
 there are also mobile and embedded computing
devices with wireless connections

08.05.25
  50
  three requirements for pervasive applications
 embrace contextual changes: a device is aware
that its environment may change all the time
 encourage ad hoc composition: devices are used
in different ways by different users
 recognize sharing as the default: devices join a
system to access or provide information
 examples of pervasive systems
 Home Systems
 Electronic Health Care Systems
 Sensor Networks

read pages 27 – 30 about pervasive computing

08.05.25
  51