Networking: Vision and Packet Switching 1959 - 1968
Intergalactic Vision to Arpanet
In October 1962, when Dr. J.C.R. Licklider became the first Director of the Information Processing Techniques Office of the Advanced Research Projects Agency, he already had a reputation as prompting leading-edge computer research. He advocated the development of time-sharing (many users using one computer at the same time) when computers of the day processed one task for one user at a time. Then in 1963 he proposed an even grander vision: an Intergalactic Network. Thousands of computers, millions of computers, interconnected so users on one computer could share information and work collaboratively with users on many other computers. Who would have ever imagined such a thought? Licklider did, and he did not hide it under a basket, but shared it with his colleagues. By dint of his energy, enthusiasm and ability to inspire and lead others, Licklider created a founding vision for computer communications, and momentum essential to this history took hold.
In 1965, two computer scientists influenced by Licklider, Dr. Lawrence G. (Larry) Roberts and Thomas Marill, conducted an experiment to understand what it would take to interconnect two computers. It highlighted the complexity of the problem, with the obvious conclusion that circuit switching -the way the telephone network worked - was a poor match for the needs of computer communication.
The telephone network had been designed to meet the needs of voice communication. Someone hand-pecked a telephone number into an instrument, and the request traveled down a wire to a switch that checked to see if the connection could be made, and did so if it could. Then and only then could the two people talk. Many seconds, sometimes even minutes, elapsed with no one taking notice, being distracted with what they were going to say or an absent thought. To computers it was an eternity, even by the standards of the slowpoke computers of the day. But how else would one build a communication system?
Earlier, in 1959, another computer scientist, Paul Baran, asked that very same question for different reasons. By 1962 he and others working with him had documented an alternative communication system, one based on exchanging small messages, not on making circuits. To most it seemed, well, absurd. Then in 1965, a British computer scientist, Dr. Donald Davies, proposed the very same idea. Roberts would return to the idea of a message-based communication system when he took on the responsibility of building a network for computers in 1968.
Roberts had to sort out many questions, such as how would one build such a network or how would it work? There was no blueprint, no prior example to be improved upon. A few ideas, some money, not much, when combined with motivated, exceptional scientists, and a vision gave rise to the birth of computer networks. In mid-1968, a Request For Quotation was distributed to potential contractors inviting bids to build such a network.