CLIENT-SERVER COMPUTING


Azam Mahmood

Client-server concepts began to gain common acceptance in 1990. They have now become popular in commercial applications of PC’s in many countries.

Client-server computing cialis cheap is one way of combining PC’s in a network. Specifically it involves designing the network in such a way that a database is stored and maintained on the network server, which makes data available to users on their PC’s, who are treated as Clients. A high-speed mainframe, minicomputer or a PC is dedicated to this task. More specifically, the processing of the application is shared between the server and the client PC, and may be controlled by the client.

Client-server computing is a form of distributed or co-operative computing where the work on applications can be off-loaded from a host or server in a network to the employee or department where the work would be done it, were being done clerically. For example, in a payroll system, the employee database may be created, updated and maintained on the server. Programmes on the server may control the updating of the database but programmes on the client computers will be used to check the process any transactions that will update the database, except those checks that require access to the database i.e. ensuring that an employee is on file and is active.

Much of the recent popularity of client-server computing is due to the development of a common language for accessing information in database called the Structured Query Language (SQL). This has enabled database vendors to use a variety of processing platforms to store the databases, yet still make information available to a variety of client PC’s running different software packages from Lotus 1-2-3 to application-specific reporting programmes.

Although client-server computing is often associated with decision support systems and Executive Information Systems, there are no limits on the application of the concept and the concept is also found in On-line Transaction Processing systems and other systems.

The prime advantage of client-server computing is that it will make it easier for end-users to develop their own applications. Second, it is claimed that client-server computing will allow MIS department . Employees to concentrate on corporate processing needs, rather than user needs. Corporate processing needs are directed by users, who will continue to ask MIS employees to do the work for them.

The third advantage claim is that client-server computing will reduce application backlogs in the MIS departments. This is not likely to happen, even if the systems development process does become more efficient through the adaptation of client-server computing and other innovative techniques. The nature of information is that each time someone obtains a piece of information; it leads to several new questions that were not foreseen earlier. This very quest for additional information shall create greater backlogs in many companies as client-server computing becomes more popular.

The major snag in client-server computing is the higher cost of maintaining software as information needs keep changing because of unforeseen questions that get asked by the consumers or buyers and the changing environment in which system has to operate. The cost of highly skilled analysts and programmers keep increasing which cannot dispense with to maintain and change the systems.

The more likely benefits will be improved customer service, employee satisfaction and faster reaction to new marketing opportunities. The benefits may reduce costs indirectly or they could increase enterprise revenues. Client-server computing will allow PCs networked in LANs to replace mainframes and minicomputers. Client-server computing can boost the performance of existing networks, create more flexible solutions and enables users to develop more novel systems solutions. As more work is taken from the server and placed on the client computer, all of these benefits can be achieved. In addition, as more processing is done by the users, many benefits can be realized in ensuring that all data on the system is more accurate, more complete and entered more promptly. This obviously increases the value of the information abstained from the system.

When migrating to client-server, the organization must cost-justify the downsizing operation and must realize cost and productivity benefits associated with downsizing.

Although client-server hardware and software costs are much less than mainframe hardware and software costs, the ongoing expense of operations for client-server today can be much higher. It would be wise to consider all hidden and up front costs.

It is important to note that there are differences between deploying client-server in a small environment versus a large enterprise setting.

There also goes a false alarm that client-server computing will force (MIS) department to disappear. Most people don’t care what platform the application is on, they hate to do all of the work that (MIS) department used to do.

As all the major database vendors vigorously promoting client-sever computing, the trend is developing and as more people are attracted to the concept, users will benefit form the experience and knowledge of other users, just us they have with other software.

Client-server Issues Starting

*      Start with a small to medium sized project. Staff can acquire experience and skills over the time.

*      Spend a lot of time on design and requirements gathering.

*      Do not jump into it with unrealistic expectations. End-user expectations can get very high and need to be managed.

*      Do not underestimate the time required to install the necessary software utilities if the skills are not readily available.

Usage

*      Develop prototypes of potential bottlenecks first, to test data transfer times and the impact on other systems.

*      Use virtual fields and triggers.

*      Put all validation rules into a data dictionary

*      Avoid transferring too much data through communication sub-systems.

*      Do realistic tests / benchmarks before implementation and remember : I / O and communication, sub-system performances degrade when overloaded. The amount of data transferred includes the data stored on the server, the programme stored on the server and presentation data.

Draw panels / windows

*      Present minimal information

*      Present detail upon request, not by default

*      Keep windows clean and uncluttered

*      Do not rely on cursor location

*      Screen painting time is very long in many products. In some products it may take 6 seconds to paint a complex window.

*      Drawing time is directly related to the amount of Information / obj ects in the window.

Successful EIS

*      Simply buying a generic Executive information system (EIS) tool will not allow you to build a. really effective and fast client / server-based EIS over your existing decade-old database. A database designed for transaction processing is totally different to that required for an EIS.

*      The most rapid and useful EIS systems use pre-summerised data with the end user querying the summary data.

*      Summarizing the data needs a clear understanding of the client requirements and means extra design and programming effort independent of the client tool used.

*      Summarizing a lot of records takes time and consumes a vast amount of CPU power.

*      Early in the design phase include the method and cost of distributing the client software

*      Throughput rate of client data routed in a twin ax and LAN communications sub systems is not close to that of channel attached hard disk drives. The many layers between the client and server almost always means slower response to database access.

Subscribe / Share

Article by nuzhat akhter

Authors bio is coming up shortly.
nuzhat akhter tagged this post with: , , Read 913 articles by