How does a tcp connection pool distinguish between different clients

In my understanding, a tcp connection in a connection pool can be reused only if the client trying to connect to the server is the one previously created the connection (i.e.same IP, etc). So when the server tries to connect开发者_Go百科 to a specific client (i.e. fetches a connection from the pool) how does the server do it? Do I have to provide IP of the client or something?

Connection pooling is a client side technology. For example, in a .NET application the first time you create a connection and close it with connection pooling turned on, the connection is not actually closed but kept open in a pool which is a cache within the client process. Another request for a connection within the same process can return the value from the pool.