Client/Server Architecture
|
The Client/Server generation stops the process of shifting work back and forth between the center and the periphery. Each computer does some computation now. How much is done where depends on the application. Let’s take a simple database example first. A common database operation begins with a table lookup and is followed by a calculation. For example, let’s say that a company has been test marketing one of its products at a lower price in one region of the country. After a week, a marketing staff person wants to access sales data to see if it would by reasonable for the company to extend the lower price to the whole country. The company secures data about all of its retail sales through grocery stores point of sales computer systems. Through a variety of stages, this raw data is uploaded to the main company’s mainframe. The company also needs to have software that allows sales from the test region to be modeled such that it will predicts, sales in the rest of the country. Such operations requires many iterations of many values on large sets of data to come to a conclusion, it takes a long time to run on even a very fast computer.
It would be possible to do this database work in a terminal-host system. The host could do all of the searching, sorting, and calculating, and send the results to the terminal. Also this could be done in a client-file server approach. All of the data would be in the host (it means server) and all of the processing software would reside in the client.
Another solution of this is to add one unit into our network. This unit is called application server and there runs the modeling program which takes the data and cranks out its analysis. When is application server used, the communication in network is following: The mainframe sorts through its records and finds the requested data. It then copies the appropriate records and sends to the application server. When the server has finished its work it sends its data to the client, which loads in user used program.
In an even more sophisticated version of this three-tier Client/Server model, the process begins with software at the client, formatting the initial query in a way that is easily understood by the host and optimized for execution of its database software. This is important because a database query can be executed very fast or very slow, depending on how it is structured. While the optimized query might come for example from Excel itself, it is more likely in a case like this that it comes from a “front end” program written specifically for the host’s database. The front end will have knowledge of how the database is organized (record structure, indexes, etc.) and will be able to do the complex calculations needed to make sure that the query can be executed in the shortest possible time. Since a modern database, particularly one user in a three tier Client/Server environment, is likely to be relational, the query will be formatted in SQL.
The number of layers in a distributed computing environment is not fixed. A two-tiered Client/Server system is quite common for smaller systems, while three tiers system is increasingly preferred for larger ones. In three-tiered systems, the layers are client, server and host.
|
|