Internet Basics: TCP/IP and DNS
There are two things that need to be understood to get an idea of how the Internet operates. These are TCP/IP, introduced in this article, and DNS, introduced in the next.
The brief article on the Internet as a Cold War Child provides some historical background to the development of TCP/IP. This article provides an introduction to what TCP/IP is and how it works.
Intended Audience
This document is intended for those dealing with computers at home or on the job. It does not intend to provide a complete technical description of TCP/IP - which would take volumes - but to provide enough understanding to relate better to help-desk and support staff, to form a better idea of how the Internet operates, and to understand the scope the subject.
What does TCP/IP stand for?
TCP/IP stands for Transmission Control Protocol/Internet Protocol. TCP/IP is a suite of software protocols and services that control the way information flows around the Internet and between computers. It is embedded in every current-model computer system and also in the routers that direct Internet traffic around the world. Later we'll look at where in your computer you will find TCP/IP settings.
What's an IP Address?
For any information to reach you via network, your computer must have an address. Although it serves exactly the same purpose as a postal address, a computer's address is formatted as a series of digits divided by points, for example:
192.168.1.1
This is called an IP Address, meaning 'Internet Protocol Address'. Every device attached to the Internet has an IP address – not only PCs, but also PDAs, and Internet-enabled cameras and printers. This means that the pool of IP addresses must be a very large number, as there are millions of such devices with many more continually being added. Later on we'll look at issues involved with maintaining the pool of available IP numbers
The Anatomy of an IP Address
Part of the IP address signifies the network identity – that is, the network which the computer is part of. Another part of the address signifies the host identity, where 'host' means an individual device attached to the network. These two components can be thought of as analagous to a street name and a house number.
An associated group of digits comprises the subnet mask. This information is used to subdivide IP address families between varying numbers of hosts - like creating room numbers representing each computer. The Subnet Mask number usually appears as 255.255.255.0.
The combination of the IP address and subnet mask must always result in a unique ID for each device, otherwise information cannot be delivered accurately, meaning either that the network will fail, or that connection to the Internet cannot be made or will perform erratically.
Now if your computers were on their own network, and you only had a small number of them, then you would only need a very small range of IP addresses to distinguish one computer from the other; you could simply number them from one to six. And early network technologies were mainly designed on this basis, as, at the time, few foresaw a day when millions of computers would be interconnected around the world. Over the years, integrating these earlier network techniques with the requirements of the Internet took some very fiddly re-engineering. However by 2004, virtually every desktop computer incorporated TCP/IP in the operating system.
Before we look at the TCP part of the protocol, we'll look at IP addresses in detail.
Incidentally, all of what follows deals only with IP version 4, which is the most commonly-encountered version at this time. IPv6 is mentioned but details are beyond the scope of this document