This document discusses distributed processing and distributed systems. It defines distributed processing as a technique that distributes information over multiple intelligent devices that are interconnected. A distributed system is one where components located at networked computers communicate only by passing messages. The document outlines different forms of distributed processing including distributed applications, distributed devices, and distributed data. It also discusses techniques like centralized, decentralized, parallel, and clustered distributed processing and challenges like heterogeneity, security, and scalability.

1. DISTRIBUTED PROCESSING
Md. Mahadi Hassan Rakib
Lecturer, Department of Computer Science and Engineering
North Western University, Khulna

2. OUTLINE
 Distributed Processing
 Distributed System
 Architecture
 Form of D.P
 Techniques
 Challenges
 Advantage/Disadvantage

3. What is Distributed Processing
 Distributed Processing is a technique of distributing the information ov
er a number of devices.
 The devices may be computers or data terminals with some level of inte
lligence.
 The devices are interconnected with communication facilities.

4. What is Distributed System
 A distributed system is one in which components l
ocated at networked computers communicate and
coordinate their actions only by passing messages.
 Examples
 The internet
 An intranet which is a portion of the internet managed by
an organization

5. Architecture
 Software layers
 System architectures
 Interfaces and objects
 Design requirements for distributed architectures

6. Software layers
 Applications, services
 Middleware
 Operating system
 Computer and network hardware

7. System architectures
 Client-server model
 Services provided by multiple servers
 Proxy servers and caches
 Peer processes

8. Design requirements for distributed architectures
 Performance issues
 Use of caching and replication
 Dependability issues

9. Design requirements for distributed architectures
 Performance issues
o Responsiveness
o Balancing computer loads
o Quality of services
 Caching and replication
o The performance issues often appear to be major obstacles to the successful deploy
ment of DS, but much progress has been made in the design of systems that overcom
e them by the use of data replication and caching.
 Dependability issues
o Correctness
o Security
o Fault tolerance

10. Form of Distributed Processing
 Distributed Applications
 Distributed Devices
 Network Management and Control
 Distributed Data

11. Distributed Applications
 One application splits up into components that are
dispersed among a number of machines
 One application replicated on a number of machines
 A number of different applications distributed among a
number of machines
 Can be characterized by vertical or horizontal
partitioning

12. Distributed Devices
 Support a distributed set of devices that can be controlle
d by processors, e.g. ATMs or laboratory interface equi
pments
 Distribution of processing technology to various locatio
ns of the manufacturing process in factory automation

13. Techniques of Distributed Processing
 Centralized
 Decentralized
 Parallel
 Open Distributed Processing
 Clustering

14. Centralized
 Centralized Processing is done at a central location, using termi
nals that are attached to a central computer
 The central computer performs the computing functions and contr
ols the remote terminals. This type of system relies totally on the
central computer.

15. Centralized
 Example: Client/server is the most common example of centrali
zed processing, where server is controlling all the activities on th
e network.

16. Decentralized
 Computer systems in different locations. Although data may be tr
ansmitted between the computers periodically
 Example: Yahoo server is the example of the decartelized proces
sing. On each login it connects you to a different server .

17. Clustering
 A cluster is a group of individual computer systems that can be
made to appear as one computer system.
 Clustering is just one form of parallel computing.
 key points that distinguishes clustering from other is the ability to
view the cluster as either a single entity or a collection of stand-al
one systems
 For example: a cluster of web servers can appear as one large w
eb server, but at the same time, individual systems within the clus
ter can be accessed as individual systems,

18. Clustering
 Example:
Internet is an example of clustering, where different server are w
orking but they look like a single server.

19. Challenges
 Heterogeneity
 Security
 Scalability
 Failure handling
 Concurrency

20. Advantages
 Quicker response time
 By locating processing power close to user, response time is typically impro
ved. This means that the system responds rapidly to commands entered by u
sers.
 Lower costs
 Long-distance communication costs are declining at a slower rate than the c
ost of computer power
 Distributed processing can reduce the volume of data that must be transmitte
d over long-distances and thereby reduce long-distance costs.
 Improved data integrity
 High degrees of accuracy and correctness may be achieved by giving users c
ontrol over data entry and storage.

21. Advantages
 Reduced host processor costs
 The productive life of a costly mainframe can be extended by off-loading s
ome its processing tasks to other, less expensive machines
 Resource sharing
 One of the main advantages of developing microcomputer networks is beca
use they make it possible to share expensive resources such as high-speed,
color laser printers, fast data storage devices, and high-priced software pack
ages.

22. Disadvantage
 Complexities
 A lot of extra programming is required to set up a distributed system
 Network failure
 Since distributed system will be connected through network and in case of n
etwork failure non of the systems will work
 Security
 The information need to be passed between the network. And it can be track
ed and can be used for illegal purpose.
 Costly software
 Not all situations are suitable for distributed computing