Chapter 6 - Networking: Arpanet 1969-1972
6.11 NPL Network and Donald Davies 1966 - 1971
When Donald Davies became Superintendent of the Computer Science Division of the National Physics Laboratory in August 1966, he finally had the budgetary power to bring into being a network exploring the packet switching ideas for which he had been proselytizing. Davies would lead a very different effort than did Roberts. He explains:
We realized that we could never build an Arpanet. We had no mandate for building a nationwide network, and couldn’t imagine who would use it. We had nothing like ARPA to act as our community.
In searching for network design objectives, his team realized they had to look no further than a problem they confronted every day in their laboratory: How to provide common computer access for their growing proliferation of computer terminals? Their solution, a computer called an interface processor, was essentially a terminal multiplexer much like the TIPs of Arpanet. Only the NPL interface processor connected up to 512 terminals to a local computer at a combined data rate of up to one megabit per second.11
Once the interface processor had been conceived, the NPL researchers questioned whether there might be more to gain than simply terminal connections. For as with all new “black boxes,” their existence opens up new possibilities, possibilities not contemplated in their original design. In the case of the interface processors, Davies’ group recognized that they could share more than just common terminal access. Davies remembers:
The main thing we found about small computers in those days, was that they had very good processing by the standards of the day, but they were very expensive when they came to adding some storage. It was expensive, because you had to add disk stores, which in those days were great big cabinets that cost you more than the computer itself. So what we said was that what we can do for all our mass of small minicomputers, PDP-8s and things around the laboratory, was provide them with a central storage facility, a file server, using the latest technology. So we built a file server to test this network.
In a paper delivered at an IFIP conference in Amsterdam in 1968, Davies discussed for the first time the concept of “local area networks;” the need for “local” computer-to-computer and terminal-to-computer interconnection. Again history will show that Davies had coined a phrase, much as he had earlier with “packet switching.”
When the network first worked in 1971, the simple host-to-host protocol proved inadequate and had to be completely re-written in order to speed up packet transmission. Once the new protocol software was installed in 1973, the network handled over one million packets a day.
Terminal addresses were organized hierarchically with three levels of branching and eight nodes at each branch; hence the address limitation of 512 devices.