2. Internet Basics

- How has the Internet developed?
- Internet Structure
- The World Wide Web

When most people think of the Internet, they think of a computer screen with pretty text and images- the World Wide Web page we are so used to. But the Internet is nor the same as a web page. It predates the Web, and has many other functions. What with email, news groups, videoconferences, chats, voice over IP, p2p networks, etc (none of these are web based), the list is huge and getting bigger.

How has the Internet developed?

The internet began in the USA during the Cold War. When the Soviet Union placed the Sputnik satellite into space in 1957, the US military was shocked, which led to the creation of the Advanced Research Projects Agency (ARPA). One of ARPA’s projects was to investigate large-scale networking of computers, especially to allow collaboration between scientists and researchers. It was to be a decentralised network, without a central focus or control point, so that if one node in the network failed, the others could take over its role. The decentralised nature of the intetrnet is said to have been inspired by the idea of surviving a nuclear attack. A centralised communication system could be put out of action easily with a missile, but if all the points on the network could replace the destroyed node, if none was essential to the functioning of the network as a whole, then the system would be able to withstand at least limited damage. In fact, the initial motivation was more how to avoid centralised control, and the need to stimulate cooperation amongst researchers. And it was to be open to connection with other systems via publicly available standards.
In 1969, ARPA scientists connected two (and soon thereafter, four) mainframe computers (there were no PCs then) in different states of the USA, and began to send data to each other via a rudimentary packet system, the beginnings of the internet protocols that we use now.

But what they were beginning to share was the fundamental basis of the internet, the TCP and IP protocols. Formally adopted in 1982, these technical standards allow the internet to function as a decentralised network of interconnected computers. The TCP protocol defines the way data is broken down into manageable chunks, or packets, which are then sent individually through the internet. Imagine a large letter is broken up into separate pages. Each page is put into its own envelope and then into the post box. When they reach their common destination, they are put back together to create the original file or piece of data. This is more efficient than, for example, a continuous flow of data, because if one of the packages is lost, it can be re-requested and sent without having to re-send all the data, causing less errors and less wasted time.

Source: http://www.asc.upenn.edu/usr/chunter/agora_uses/chapter_2.html

The IP protocol controls the way the packets reach their destination, a kind of addressing system based on IP numbers such as 123.123.123.123. It is a bit like the address on an envelope. Each computer on the route to the destination knows how to send the packet to its IP destination, choosing the best route according to which computers are available and connected at each point in time. There is no direct connection between the origin and the destination, and the route is not decided in advance before the journey begins. Routers decide where to send the packets, depending on which other hosts are available, and the packets jump from one host to another until they finally reach their destination. To be on the internet, you need an IP number, as well as a physical connection such as a phone line, an Ethernet or cable connection. In theory, no IP number is more important than another – the internet has no central brain or headquarters. In practice, as we will see, some IP numbers, such as those assigned to the DNS databases, routers and backbone nodes, are more important than others.

Illustration 3

This system is different from the traditional information flow used in the telephone system, called circuit switching. Here, the flow of information is continuous, along one channel, and if this circuit is broken the connection is lost and communication stops. But in packet switching, if the flow is stopped for some reason, then the routers are able to find alternative routes, and not all packets need to follow the same routes as each other before they are brought together and reassembled at their destination.

By 1984, the number of IPs was around a thousand, and it was impossible to remember them all, so the Domain Name System (DNS) was introduced. This meant that an internet address could be made up of words, not just numbers. When you typed in the address, whether in email, or a news group (remember, at that stage there was still no WWW or hypertext), your programme asked a central database for the IP number that corresponded to that name, and used that IP number to send the packets to their destination. For a name to work it had to be registered first; otherwise it would not be in the data-base. All internet name addresses (as distinct from IP addresses) must use a domain. The first domains were .mil, .edu, .com, .org, and country domains such as .uk were introduced from 1985. That same year, the task of running the DNS database was given to the Information Sciences Institute (ISI) at University of Southern California (USC), and domain name registration was run by the Stanford Research Institute (SRI). The internet was still very much a network for researchers, and the organisations that ran it were university-based. In 1986, the National Science Foundation set up NFSNet, which provided a backbone of fast fibre optic connections to which other centres in the USA, mostly universities, could connect easily. The Internet Engineering Task Force was set up the same year to decide on the technical standards for the entire internet also with a little US university funding. This means that the US government essentially paid for the physical infrastructure and the running of the en-tire network, with the exception of those networks in other countries which were beginning to connect up to the US network, and which were research networks funded in turn by their own governments. There was still no commercial internet, and almost all users were from universities or research institutes. It was not until 1993 that the introduction of the graphical World Wide Web made it much easier to present and find information on the internet. Together with the commercial exploitation of the internet, which took off around 1994, this created the expansion boom that has led to the millions of users we have now.

Internet Structure

How does the internet work now? The underlying basis is the same: a whole lot of computers connected up to transfer data from one to the other. Data can flow through fibre optic cables, copper wires, coaxial cables, copper telephone lines, satellite connections, wireless, Ethernet cables, etc. It still uses the TCP/IP protocol that moves the data from place to place. On top of this, there are various other protocols that allow us to do useful things with the data. For example, the SMTP protocol sends email on its way to another internet server. It is not necessary to use the SMTP server on your own ISP system, although most systems are protected from external spammers so that only that ISP’s users can access the data. The HTTP protocol allows your web browser to send a request for one or more web files to another computer (web server). When the text and graphics files reach your browser, it interprets the HTML language in the text files, laces the graphics and the colour in the appropriate place, and assembles them all to create the visually-pleasing page you see on your screen. The POP protocol allows us to login to a POP server and download our mail to be stored on our own computer, rather than only view it on a web page while we are connected to the internet. IMAP allows many other kinds of files and applications, such as audio and video, to function together via the web.

To connect to the internet, you can be on a local network which is connected already, as in a library or university, or have a special cable which connects you directly, such as a leased line (a special cable leased from the telecommunications company which provides fast but expensive access) or a cable TV connection, or have a contract with a company which provides a satellite connection, or through a local wireless connection point. Most people however, connect via the telephone system with a modem or ADSL, using their telephone wire (via the ‘local loop’ between the telephone and the telephone exchange) because it is cheaper and more readily available. In general, the faster the connection, the more it costs.

When you connect via a commercial link, it is normal to do that through a company or Internet Service Provider (ISP), which has a permanent connection to the internet and charges you for sharing this connection with hundreds or thousands of other users. The ISP only routes local traffic to a commercial carrier, usually a telecommunications company or a huge ISP, with a further connection to the main highways of the internet, which most people never see. The first of these backbone carriers was NFSNet, back in 1986, but now huge companies, such as UUNet, Sprint, and ATT, run the most important internet infrastructure. They connect to each other through Major Exchange Points (MAPs, MAEs, etc) and in this way the internet extends over the face of the planet. So we have:
1. End users
2. ISPs
3. Carriers
4. Major Exchange Points
5. Backbone carriers

Most of this infrastructure is owned and managed by private enterprise, not, for example, by governments, and hardly ever by community organisations. This early privatisation has had important consequences for the future development of the Internet, and this is explored in more detail in chapters 3 and 4.

The World Wide Web

The early internet had no graphics. Everything was in text form: letters, numbers and symbols on the screen. The introduction of the World Wide Web (‘the Web’) revolutionised the internet, making it more attractive, more versatile and easier to use. The Web was initially conceptualised by Tim Berners-Lee and other scientists at the European Centre for High Energy Physics (CERN) in Geneva, Switzerland in 1989, and a simple – and free
– browser was released to the public a year later. Initially progress was slow, with no more than 150 web sites in the world by the end of 1993. It was the creation of the Mosaic browser programme by Mark Andreeson in 1993 that simplified enormously the use of the Web and made the pages so easy to read and visually pleasing. This was the predecessor of Netscape and Microsoft Explorer. The programme was made available to the public, especially to the educational community, and it rapidly replaced the text-based tools for information retrieval like Gopher, Archie and Veronica that had been used before. In 1994, the WWW edged out telnet to become the second most popular service on the Net (behind ftp-data) based on the percent of packets and bytes traffic distribution on NSFNET, and a year later it became – and has remained
– number one.

It was in 1994 that private enterprise started using the internet, with the commercial sites and the first virtual shopping malls and cyberbanks. In 1995, NFSNet became a research network again, and the infrastructure was now firmly in the hands of private enterprise. The internet began to be sold as a commodity, and it sold well. In 1994, there were 3,000 web sites and a year later 25,000.

There are many problems with this way of financing internet infrastructure. To begin with, monopolies in the telecommunications industry can mean that prices are artificially high and competition is minimal. The USA dominates the internet, both in terms of number of users and amount of content. Economies of scale mean that it is cheaper to connect to the internet in the USA than in other parts of the world, since the infrastructure there is better developed and US companies control it worldwide. And, because of the way ISPs are charged for their connectivity, it is more expensive for ISPs in poor countries than in rich ones. Coupled with higher fees for telephones, the result is that users in poor countries have to pay much higher prices than those who can afford to pay more. This means that very few can benefit from the advantages of the internet in those countries. Internet infrastructure used to be horizontally organised, but now, with the dominance of huge multinationals, it is much more vertically organised.¹

Source: ITU, OCD, World Bank

1 Dogan, P, Vertical Relations and Connectivity in the Internet', in Communications and Strategies, 47, 2002, pp87-101