How the Internet Works
Some of the inspiration for the Internet came from academic research into networks that could survive a nuclear attack. As a result, one of its most distinctive features is its lack of centralised control. Instead of one big computer that controls everything, the Internet is rather ramshackle network of all kinds of different computers. It only works at all because everyone who uses it agrees to use the same technology.
Once you start using the Internet, you will notice that information seems to travel across it in fits and starts. In fact, information travels in packets – small chunks of information that act independently of each other. Packets can go across the Internet via any available route. The real-world equivalent would be a newspaper delivery service that sends one page via the next street, another via the next town, another through a different country and another to the moon and back, before finally reassembling the newspaper and pushing it through your letterbox.
This may seem wasteful, but it allows the Internet to deal with problems intelligently. If one computer fails, packets are sent over a route that avoids it. This switching process happens instantly, and no information is lost.
The Internet’s other great strength is that the system that controls packets – called TCP/IP (Transmission Control Protocol / Internet Protocol) – works in exactly the same way on every computer. TCP/IP is powerful enough to do the job of creating packets and reassembling them when they arrive – even if they arrive out of sequence – but simple and undemanding enough to work on small personal computers as well as large scientific and commercial ones.
Personal computers use a small piece of software called a ‘TCP/IP stack’ to do this job. This is supplied with some Internet software, but in many cases it is included with your computer when you buy it. This provides that pipe that makes information flow in both directions between the personal computer and the Internet. If the stack is not working, the Internet connection will be dead. When you buy a connection from an ISP, what you are really paying for is the technology needed to link this pipe to the ISP’s own network, which has a permanent direct connection to the Internet. All the software you use relies on the stack being in place and working correctly.
On the Internet, computers and their software are often divided into three categories. A server is a computer system that offers access to a particular service or facility, usually on the Internet. A mail server collects and stores all the mail that arrives addressed to an ISP’s customers and a news server does the same for news. A Web server serves up Web pages on the Web so that they can be read by anyone. A news server serves up copies of newsgroup postings and so on.
Although servers are usually big, fast and powerful computers that can handle thousands of requests for information every minute, it is possible to install server software on any computer. one common situation in which this is necessary is when someone is developing a website on a home PC. Pages can be viewed with a browser while they ate being developed, but without a personal Web server installed on the home computer none of the links between pages in the site will work. This makes it hard to check that the links are working properly. With a server, all the links within the site will work correctly because it will be able to serve up the relevant pages when asked. This means that you can test the site in rpivate at home before making it public by copying it to an ISP’s Web server.
Clients are computer systems that copy information from a server to you and display it in a form that you can use. An email client connects to the mail server, finds your messages out of the thousands of others that are stored there, copies them to your computer and organises and displays them for you so you can read and reply to them. In practice, most Internet software used on home computers is client software, even though it is not always labelled as such.
The client/server idea is an important one and may have a big effect on the way that the Internet develops in future. One proposal from the past was the thin client – a computer system that is not much more than a screen and keyboard, with some memory, a processor chip to do the work and perhaps a hard disk. The thin client tried to move the maintenance usually associated with personal computers – especially the frustration which software installation can sometimes cause – over to a big general –purpose server computer. Instead of installing the same software on many machines, it needs to be done once only.
You will sometimes see the words server-side and client-side. These are industry jargon for used by servers and used by clients. For example, server-side software is simply software that can be used on a server.
Hosts are more or less the same as servers. A host computer is one that hosts some action or exchange of information. In most contexts, the word simply means “the computer at the other end of the link”.
