A Brief History: The Internet

The Internet originated at the Defense Advanced Research Projects Agency (DARPA) in the USA around 1962. The original concept, and one which is an integral part of today's Internet, is a way to share a single communication line with all those connected being able to talk at the same time. Imagine a conference where everybody is speaking together. It usually takes ages to call the house to order. However, a computer scientist at the Massacheusetts Institute of Technology invented the idea of packet switching. Imagine now that everybody's conversation at the conference is chopped up into snippets and each person takes it in turn to utter a snippet. The clever bit is having each listener taking in each snippet aimed at them and reassembling the snippets in the right order so they can then hear the others' entire conversation. Obviously this doesn't work too well with humans, but it does work with computers, except the snippets are called packets. Each packet contains some of the data, together with the address of the computer it is destined for, together with a sequence number so that the packets can be reassembled in the correct order.

If you can conceive of the Internet as a single, long piece of wire, with every computer connected to it. You are not too far away. When a computer wants to communicate with another computer it looks up the address, and then sends the data as a series of electronic pulses along the wire, rather like the sound made by a fax machine but using electrical waves rather than sound waves. Now, more than one computer cannot send information one one line at the same line because the signals would get mixed together and make nonsense of the message. To make it fair, the computer listens to the line and, if it is not busy, proceeds to send, or broadcast, a packet. If the line is busy, the computer waits a short time and tries again. If, after listening, two computers manage to broadcast simultaneously, the mistake is recognised and both back off for a short but random amount of time, to minimise further clashes. This scheme has two great advantages: firstly there only needs to be one wire, for any number of computers connected to it, instead of a separate wire linking each pair of computers, and secondly sharing the line is very fair. Rather than a computer queuing up to send a short e-mail behind a computer that was sending a huge report down the same line, the few packets from the first computer get sent in-between those from the second.

This sounds like an awful lot of work and indeed it was. Data was first sent at around 75 bits per second. If you tried to download a 6 megabyte song from the Internet, if there was such a thing forty-two years ago, it would have taken 148 hours, or over six days. As we now know, things progressed, computers got faster, computer scientists developed compression to reduce the size of messages, and therefore increase the speed of their transmission.

There was, however, one further problem to solve. As more computers are connected to a single wire they spend more time listening and waiting for a quite period than they do transmitting data. This placed a practical limit on the number of computers. The solution was to separate the wire into segments, and place a special type of computer, known as a router, between the segments. If the router saw a packet one one segment that was destined for a computer on another segment, then is passed it across, otherwise the packet remained on its original segment. That way each segment had less of the overall traffic. The final trick was to arrange to have computers that communicated with each other frequently connected to the same segment, whilst rarer communicators were put on separate segments.

This is ancient history. It was not until 1969 that the Advanced Research Projects Agency built the first network and called it, rather imaginatively, ARPANET. In 1970 researchers finalised a protocol, known as NCP or Network Control Protocol. that standardized computer-to-computer communication. In 1972 the forst "hot" application, e-mail, was created. On the first of January everybody connected to the Internet at that point changed over from NCP to TCP/IP , which is the standard still used today. Ten years of dedication, vision, experimentation and belief of researchers, scientists and engineers was all that was needed to fuel that which has become the Internet that we see today.

The World-Wide Web

Today
The construction of a network is known as its topology ,and below are some topologies that are commonly found in homes, businesses, global corporations and, indeed, the Internet itself. Simply by drawing further lines between each illustration it would be easy to see how complicated things become.

Here we see a simple network in which all the blue computers are connected together by a single wire.

Fig. 1. A basic network.

A hardware device called a hub is used to simplify cabling. It is generally easier to run single cables to a central point than to run a single cable around a whole office or building. It also prevents a lone cable from being a single point of failure.

Fig. 2. A basic network with a hub instead of a single wire.

The router minimises traffic across the whole network by filtering packets that are not addressed to another segment.

Fig. 3. Two networks connected by a router.

Bridges are often used to connect smaller networks that are separated by some distance. In fact the connection between the devices at either end could be via satellite, fibre-optic cable or a transatlantic link. Modern bridges also act as routers to limit traffic on the bridge itself.

Fig. 4. Two networks connected by a bridge.

Network topologies usually fall into two types: a Local Area Network, or LAN , and a Wide Area Network, or WAN . The difference is usually the geographical spread of what someone considers to be a single network under their control. Figures 1. and 2. above are obviously LANs, but Figures 3. and 4. could be LANs or WANs. A large company with several buildings located nearby could use bridges to link them, but still consider the whole thing to be a LAN. A WAN might link different parts of the country and travel at least part of the way along communication lines that are under the control of a third party. In many senses, the Internet is an unmanaged network.

Fig. 5. Two networks connected by an unmanaged network.

However, the term "unmanaged network" is something of a misnomer. The Internet is managed, in many important respects, but it is not necessarily under the control of the people on either end. For this, companies can and do lease private, often high-speed, lines between remote locations.

finally, a router combined with a Firewall provides a break between a local network and the wider world. Most devices, except the very cheapest, also offer a DMZ or de-militarized zone. This allows the network to be divided into three sections: the local network and the wider network, with a firewall between them, and a third section, the DMZ, that can be seen by the other two. This is useful for things like e-mail and web servers that are visible to the public and can be maintained from the local network.

Fig. 6. Firewall/Router With DMZ.

Privacy

Another aspect of the Internet is lack of privacy. Users send packets of data down communication lines and that data is seen by other computers on the network. Ideally, the other computers, except for the real target, would ignore them, but it is possible for an unscrupulous hacker to look at the data. The simplest way is to encode it using a method known only to the sender and the recipient. The sender encodes the data before sending it and the recipient decodes it before reading it. A typical schoolboy method of encoding is shown in Fig 7. The alphabet is transposed two letters to the right so that A becomes C, B becomes D, etc. This is known as the "key". So when the sender chooses to write the word HACKERS, it is encoded to read JCEMGTU. Not knowing the encoding makes the result obscure. Of course, this type of thing is very easy to break and so we turn to encryption using good mathematics.

Encryption is about using mathematical functions that are very hard to break. In fact, some that are in use today are, to all intents and purposes, unbreakable. All the fastest computers in the world could not crack the codes given thousands of years to try. It has so far been impossible to break the codes mathematically, so the attack must use "brute force"; trying every possible combination of key. Of course, the attacker might get lucky by finding the correct key straight away, but this is extremely ulikely.