Multiplexing TCP Connections #

Socket APIs are blocking by default. This means we need to add some facilities to make them work with multiple concurrent connections to have a smooth user experience.

Polling Non-Blocking Sockets #

It is possible to configure sockets to use a non-blocking operation. One way to do this is by using the fcntl() system call with the O_NONBLOCK flag, although other ways also exists.

Once in non-blocking mode, a call to recv() with no data will return immediately.

Polling is the act of continuously running a function over a set of time. This usually wastes resources as most of the time there is no data to be read or write.

Forking and Multithreading #

Forking creates a new thread or process that handles a block of execution. This can be used to service multiple clients in a socket connection as they only block the thread or process that they're servicing.

This is useful but has some downsides:

• Threading is hard, especially if connections share any state between them.
• It is not portable as Operating Systems provide different APIs for this feature.

The select() function #

The select() function receives a set of sockets and tells which ones are ready to be read.

It can also tell which sockets are ready to write to and which sockets have exceptions.

This is also portable.

Synchronous multiplexing with select() #

The select() function has several useful features.

Given a set of sockets, it can be used to block until any of the sockets in that set is ready to be read from.

It can also be configured to return if a socket is ready to be written to or if a socket has an error.

Additionally, select() can be configured to return after a specified time if none of these events take place.

| For more info, open the man page: man select