C# Network Programming Classes

Summary

This chapter covers a lot of ground in discussing the history of network programming and its relationships to C# network programming. Network programming got its real push in the early days of Unix systems with the Berkeley socket library. Version 4.2 of the BSD Unix operating system included a common socket interface to assist programmers in communicating with other systems across a network.

The Windows socket interface, Winsock, builds on the existing Unix network socket programming model, adding many Winsock-specific functions to enhance the standard socket programming model. Many of the enhancements center around adding the capability for asynchronous socket I/O. The WSAAsyncSelect() and WSAEventSelect() functions enable Winsock programs to use standard Windows messages and events to signal socket events.

The .NET Framework offers a socket interface to access the Winsock APIs. The System.Net_.Sockets namespace provides the Socket class that uses managed code to provide Winsock functionality to C# programs. In addition to the normal Socket class, C# also offers the TcpClient, TcpListener, and UdpClient classes. These classes take care of much of the programming overhead associated with creating sockets, and they provide an easy interface for novice network programmers to work from.

As you can see, network addresses are crucial to socket programming. Unfortunately, these days there are lots of options for how to reference an address on the Internet (or even a local network). The next chapter describes how to use the Domain Name Service (DNS) classes in C# to resolve hostnames into IP addresses for your network programs