I wanted to solve the Defense Department’s problem: it did not want proprietary networking and it didn’t want to be confined to a single network technology. Vint Cerf
The implementation of the early ARPANET attracted some significant individuals. When UCLA began to prepare for the receipt of its first IMP, graduate student Mike Wingfield stepped up to build the interface to connect the IMP with the university’s Sigma 7 computer.
Steve Crocker and Vint Cerf wrote the necessary software, having been involved in the submission of one of the IMP bidding companies. Cerf earned a mathematics degree at Stanford, joined IBM briefly, and then went to UCLA for his master’s and doctorate. There he met Bob Kahn who was working on the hardware side of ARPANET.
In the early 70s, Cerf who was an assistant professor at Stanford and his students developed the core protocol for the ARPANET to manage its messages as a network of networks. In 1972 the International Network Working Group was formed and chaired by Cerf to consider how further networking technologies might be advanced in various countries across a variety of different networks.
Cerf collaborated throughout with Bob Kahn who was on the East Coast having a spell at Bell Labs and a professorship at MIT AND working on IMPs at BBN.
While at IPTO Kahn recognised that the emerging plethora of networks was going to become an issue. He collaborated with Cerf to develop a means whereby the networks might be interconnected with the host computers being responsible for their own reliability, rather than it all coming back to the network operator.
The approach owed a great deal to Pouzin’s Cyclades in that the router would not be concerned with ensuring messages were accurately and timely delivered – that was the host computer’s obligation.
The duo published a paper in May 1974 entitled ‘A Protocol for Packet Network Interconnection’ which was described as an internetworking protocol using packet-switching among the nodes to share resources. This defined TCP/IP. Transmission Control Protocol required the development of Internet Protocol to handle the packets by encapsulating them in an envelope for routers to pass around. The May 1974 paper described how the combination of TCP and IP could form an internet protocol suite. This fundamental software could then have other protocols built around them as a stable base.
Now the ARPANET had grown up and was ready to birth the Internet. Through the late ‘70s/early ‘80s Cerf worked on TCP/IP technologies with DARPA providing funding.
Cerf and Kahn later combined to found the Internet Society – to provide leadership in Internet related standards, education and policy.
Back in 1963 Lick Licklider of ARPA had defined an intergalactic computer network and addressed his audience as ‘members and affiliates of the intergalactic network’.
In the 1990s Cerf worked with NASA’s Jet Propulsion Laboratory to develop the Interplanetary Internet.
This foresaw an Internet for planet to planet communication, the large distances between planets requiring a different protocol to deal with the delays and therefore higher incidence of errors.
The Internet was to become the ‘network of networks’ but the Interplanetary Internet was planned as the ‘network of Internets’ In 1998 Cerf worked on InterPlaNet for NASA, intending to operate first with the Mars Telecommunications Orbiter from 2005, but this project was cancelled.