Multithreaded server overview
The chief advantage of a multithreaded server is that long-running client requests will not block the server from accepting other client requests. If a new thread is not created, then the current request will be processed. It is only after the request has been processed that new requests can be accepted. Using a separate thread for a request means that connections and their associated requests can be processed concurrently.
When using a multithreaded server, there are several of ways of configuring the threads as follows:
Thread-per-request
Thread-per-connection
Thread-per-object
In the thread-per-request model, each request that arrives at the server is assigned a new thread. While this is a simple approach, it can result in the creation of a large number of threads. In addition, each request will often mean that a new connection will be created.
This model works nicely in an environment where the previous client request does not need to be retained. For example...
