Client/server is two programs talking to each other (see Figure below).
Figure: The essence of client/server.
Here we see Program 1 asking Program 2 for some information. Program 1 is the client and Program 2 is the server. Program 2 serves Program 1 with the information it requested. This is different than a main program calling a subroutine and returning. A program that calls a subroutine transfers control to the subroutine and cannot perform any processing until the subroutine returns control.
With client/server, the client and server programs are independent processes. If the client sends a request to the server, it is free to perform other work while waiting for the response.
Figure below: shows the three standard client/server configurations. R/3 can be tailored to run in any of these configurations.
Figure: One-, two-, and three-tiered client/server configurations.
When the client and server programs both run on the same computer, the configuration is referred to as single-tier client/server. (A tier is the boundary between two computers.) When they run on different computers, the configuration is referred to as two-tier client/server.
A program can function as both a client and a server if it both requests information and replies to requests. When you have three programs in communication, such as is shown in Figure 1.16, the configuration is called three-tier client/server.
The client/server configuration enables the R/3 system to spread its load across multiple computers. This provides the customer with the ability to scale the processing power of the system up or down by simply adding another computer to an existing configuration, instead of replacing a single computer that performs all of the processing, such as that which occurs in the mainframe world.
No comments:
Post a Comment