Securing a computer is easy: Disconnect it from all networks, remove all input and output, and lock it in a bunker. Oh, wait—did you want the system to do something? Then you’ll probably want to connect your system to the Internet.

Many system administrators have a vague familiarity with the basics of networking, but to be a truly competent sysadmin, you need a real understanding of how everything fits together. You don’t need to know when to use rapid spanning trees, how to choose between BGP and OSPF, or even what those acronyms represent. But you must know what an IP address is, how a netmask works, how port numbers differ from protocol numbers, and why you cannot use telnet(1) to test UDP connectivity. Without this basic knowledge, you’ll fumble. Read this chapter and understand it, and you’ll have an easier time convincing your network administrator to give you what you need.

While this chapter offers an overview of TCP/IP, it doesn’t cover the innumerable details, caveats, annoyances, peccadilloes, and blatant outrages present in the protocol. If you find that you need to torture yourself with the finer points of TCP/IP, pick up one of the big, thick books on the subject. The TCP/IP Guide by Charles M. Kozierok (No Starch Press, 2005) is an excellent place to start.

This chapter covers both TCP/IP version 4 (the Internet protocol widely used for the last 30-odd years) and the new version of the protocol, TCP/IP version 6. Despite the different version numbers, the two protocols are more similar than not.

We’ll start with the layers of the network and then delve into how the protocols work.