Chapter 3 :
Distributed Data Processing
Business Data Communications, 4e
Centralized Data Processing
Centralized computers, processing, data,
control, support
What are the advantages?
Economies of scale (equipment and personnel)
Lack of duplication
Ease in enforcing standards, security
Distributed Data Processing
Computers are dispersed throughout
organization
Allows greater flexibility in meeting
individual needs
More redundancy
More autonomy
Why is DDP Increasing?
Dramatically reduced workstation costs
Improved user interfaces and desktop power
Ability to share data across multiple servers
DDP Pros & Cons
There are no “one-size-fits-all” solutions
Key issues
How does it affect end-users?
How does it affect management?
How does it affect productivity?
How does it affect bottom-line?
Benefits of DDP
Responsiveness
Availability
Correspondence to Org.
Patterns
Resource Sharing
Incremental Growth
Increased User
Involvement & Control
End-user Productivity
Distance & location
independence
Privacy and security
Vendor independence
Flexibility
Drawbacks
DDP
More difficulty
test & failureof
diagnosis
More components and dependence on
communication means more points of failure
Incompatibility of components
Incompatibility of data
More complex management & control
Difficulty controlling information resources
Suboptimal procurement
Duplication of effort
Need for new
applications
Reasons
for
DDP
On large centralized systems, development can
take years
On small distributed systems, development can be
component-based and very fast
Need for short response time
Centralized systems result in contention among
users and processes
Distributed systems provide dedicated resources
The DP “Pendulum”
Centralized systems (mainframes, etc)
Distributed systems (PCs)
Networked systems
Client-Server computing
Client/Server Architecture
Combines advantages of distributed and
centralized computing
Cost-effective, achieves economies of scale
Flexible, scalable approach
Intranets
Uses Internet-based standards & TCP/IP
Content is accessible only to internal users
A specialized form of client/server
architecture
Extranets
Similar to intranet, but provides access to
controlled number of outside users
Vendors/suppliers
Customers
Distributed applications
Horizontal partitioning
Different applications on different systems
One application replicated on systems
Example: Office automation
Vertical partitioning
One application dispersed among systems
Example: Retail chain POS, inventory, analysis
Distributed data
Centralized database
Pro: No duplication of data
Con: Contention for access
Replicated database
Pro: No contention
Con: High storage and data reorg/update costs
Partitioned database
Pro: No duplication, limited contention
Con: Ad hoc reports more difficult to assemble
Networking Implications
Connectivity requirements
What links between components are necessary?
Availability requirements
Percentage of time application or data is available
to users
Performance requirements
Response time requirements