Explain the Architecting For Threads
When available, threads are an integral part of any multitasking server application program. It is important that the operating system provide an efficient mechanism for thread creation, execution, and manipulation. Ii is this increased efficiency and decreased overhead that makes the threads an invaluable resource. Threads may be widely used to make the server application more concurrent. Thread may be widely used to make the server application more concurrent. Introducing more executing agents working on behalf of the client application component increases the performance. For the developer, there is an even finer level of control over the application. In a multithreads system, each application program may have more than one thread executing.
Multithreading adds power to the developer by allowing the programmer to dispatch multiple (> 1) units of execution. The application can execute more quickly and efficiently because each thread is working concurrently with the others. In may example of multitasking, I showed how the operating system increased efficiency by running multiple applications at the same time. Application developers can add that power to their server applications by starting multiple threads. With each thread executing portions of code concurrently, application throughput is increased. Architecting code to fully utilize threads will maximize the efficiency by using the available CPU cycles. Server application developers should take this into consideration for optimizing their code on a given platform.
A common mechanism for architecting server applications using threads is the Worker Model discussed in the next section. It should be noted that resource requirements in the Worker Model may be relatively high. For example, stack allocation of 8 K FOR 250 threads alone would equal 2MB. Other designs, modeled after one thread per client, are acceptable as well. It is possible to define pools of available threads to execute application requests as they are received.