Back to top

Chapter 08 - Networking: Diffusion 1972-1979

8.0 Overview

The successful demonstration of Arpanet at the International Conference on Computer Communications (ICCC) in October 1972 proved a turning point in the history of computer communications. There did remain much to do before Arpanet functioned as envisioned by its creators, work that would continue under the auspices of the IPTO. But a working network also presented a conundrum: neither IPTO nor DARPA had charter authority to operate a network. It had to be transferred to a private organization. AT&T exhibited no interest and the other most likely organization, BBN, seemed equally disinterested. That was until some of its key employees resigned to start just such a company. With its hand forced, BBN hired Larry Roberts to commercialize the Arpanet technology.

A working Arpanet also encouraged the IPTO to build a packet radio network. To lead the project, they recruited Robert Kahn from BBN. He immediately tackled both the network software problems of Arpanet and the issue of how to interconnect Arpanet to a packet radio network. Working with Vinton Cerf, they designed an entirely new network software protocol: Transmission Control Protocol (TCP).

The Europeans, not to be left behind, launched a number of networking projects about the time of the ICCC. In addition to creating networks very different than Arpanet, including pure datagram networks, the Europeans led the creation of an International Networking Group that then became International Federation of Information Processing (IFIP) Working Group (WG) 6.1. For those debating the nature of computer communications, IFIP WG 6.1 became a maelstrom. The gulf between data and voice communications seemed unbridgeable. The computer researchers sought harbor in standards making bodies and instigating new networking protocols known as Open System Interconnection.

Meanwhile the proliferation of computers in the United States stimulated creation of many new networks. David Farber led the innovation of a token ring network interconnecting many computers and peripherals using twisted pair wire.

At Xerox PARC, Robert Metcalfe creatively adopted many of the ideas of the Arpanet and the ALOHAnet to create another type of local area network. Ethernet let computers broadcast their packets whenever they wanted as in ALOHAnet - no passing of permission tokens like token ring. Etherent worked magnificently and entailed the design of yet another networking protocol: PARC Universal Packet (Pup). Pup would in turn be redesigned into Xerox Network System (XNS) to accommodate an even faster Etherent.

By the end of the 1970’s, the work initiated with the packet radio network project, clarified in the IFIP WG 6.1 meetings, and intensified with the innovation of local area networking led not to a unified approach for computer communications but to three computer networking protocols and many networks. How would sense be made of this chaos.