The send() and recv() call conversation

Client and server communicate using send() and recv() as shown below:

num = send(s, addr_of_data, len_of_data, 0);
num = recv(s, addr_of_buffer, len_of_buffer, 0);

The send() and recv() calls specify:

The socket s on which to communicate
The address in storage of the buffer that contains, or will contain, the data (addr_of_data, addr_of_buffer)
The size of this buffer (len_of_data, len_of_buffer)
A flag that tells how the data is to be sent

Flag 0 tells TCP/IP to transfer the data normally. The server uses the socket that is returned from the accept() call.

These functions return the amount of data that was sent or received. Because stream sockets send and receive information in streams of data, it can take more than one send() or recv() to transfer all of the data. It is up to the client and the server to agree on some mechanism to signal that all of the data has been transferred.

When the conversation is over, both the client and the server call close() to end the connection. Close() also deallocates the socket, freeing its space in the table of connections. To end a connection with a specific client, the server closes the socket returned by accept(). If the server closes its original socket, it can no longer accept new connections, but it can still converse with the clients to which it is connected. The close() call is represented as follows:

close(s);

If you are writing a client application, you might want to verify the processes the server will use. Both client applications and the servers with which they communicate must be aware of the sequence of events each will follow.