The internet's first ARPANET links

The story

Computers that did not talk to each other

In the 1960s computers were enormous and very rare machines. They filled whole rooms, cost as much as a house and were found only in universities, laboratories and the most important offices. Every computer was an island: a researcher who wanted to use the data on another machine, far away, had to catch a plane with a magnetic tape in a suitcase, or use a different terminal for each connection. For people working in research it was a great waste of time and money.

In the United States a government agency, ARPA (the Advanced Research Projects Agency), set up in 1958 after the launch of Sputnik, funded research in the universities. One of its managers, Larry Roberts, had an idea: to connect the computers of the various research centres, so that they could exchange information and share tasks. The project was called ARPANET.

Packets of data

The problem was how to send the information. On the telephone, to talk you need a line reserved for the whole length of the call. The scientists chose a different system, studied in the 1960s by the American Paul Baran and the Welshman Donald Davies: “packet switching”. A message is broken into many small packets, each with an address; each packet chooses the clearest route by itself, and when they arrive they are put back in order, like the pieces of a jigsaw puzzle. In this way the lines are used better and, if one road is cut, the packets go by another.

To make the network work, a small company in Cambridge, Massachusetts, BBN, built special computers, called IMPs, as big as a fridge. Every connected centre had one, and it acted as a “postman” between its computer and the network.

“Lo”

The first connection was between the University of California, Los Angeles (UCLA), in Leonard Kleinrock's laboratory, and the Stanford Research Institute, more than five hundred kilometres away. On the evening of 29 October 1969 a student, Charley Kline, tried to type the word “LOGIN”. He phoned his colleagues at Stanford to check: the “L” arrived, then the “O”, and then the system crashed. So the first message on the Internet was “LO”, almost “Lo!” as in “Behold!”. After an hour the connection was fixed and the whole word arrived.

The first users were mostly researchers and university students. There were no web pages, no pictures, no search engines: only text on dark screens, typed on noisy keyboards, and connections much slower than today's, even thousands of times slower. Yet even then the network served what it still serves for: talking to each other and working together from far away.

By December 1969 the network had four nodes: UCLA, Stanford, the University of California at Santa Barbara and the University of Utah. Then it grew fast: by 1971 there were already about fifteen, and the engineer Ray Tomlinson sent the first email, choosing the “@” symbol to separate the user's name from the computer's. In 1973 the first links abroad arrived, in Norway and England.

A common language

In the meantime many different networks were being born, which could not talk to one another. In 1974 two engineers, Vint Cerf and Bob Kahn, invented a common language, called TCP/IP: a set of rules that lets different networks exchange packets. On 1 January 1983 the whole of ARPANET switched to that language: for many historians this is the true birth date of the Internet, the “network of networks”. ARPANET was switched off in 1990, but by then the Internet lived on by itself.

Italy joined the network on 30 April 1986: the first Italian connection was made from Pisa, from the CNR computing centre (CNUCE).