Robert E. Kahn
President
Corporation for National Research Initiatives

My first experience with computers, aside from a brief episode toggling the switches on an early IBM model 650, was with the IBM 704, 709, and finally 7090 at Bell Laboratories in the early 1960s. In those days we submitted computer programs to large "batch-processing" machines via punched cards, carrying them by hand to the computer center for processing later. It usually took several tries to fully debug even relatively simple programs, which usually translated into several days of elapsed time from start to finish.

The introduction of time-sharing procedures, which allowed several users to take turns using the same computer via teletype machines, made it possible to debug programs interactively, thus shortening the time to create working programs. Eventually we were doing this over telephone lines, at initial speeds of 300 bits per second. However, errors on the phone lines, along with the high cost of long-distance telephony and error correction, made interactive computer communications impractical over long distances.

The ARPANET, which leased 50-kilobit-per-second lines from the telephone company, was a significant improvement, allowing messages to be communicated between computers in the United States in a fraction of a second using a technique called packet switching. During the 1970s, more than 100 time-sharing machines were connected to the ARPANET and many thousands of researchers could access remote machines, do computations involving multiple machines, and even make use of specialized facilities such as the ILLIAC IV, the world's first supercomputer.

When I began working at DARPA (Defense Advanced Research Projects Agency) in the early 1970s, I recognized a need for multiple networks to communicate. I invited a colleague, Vint Cerf, to work with me in creating a logical architecture for connecting multiple networks and the computers attached to them. The key ingredients of this architecture were a protocol, now known as TCP/IP (transmission control protocol/internet protocol), with Internet addresses (i.e., IP addresses) to identify the individual machines and gateways (now known as routers) to provide linkages among networks. The networks that constituted the initial Internet were the ARPANET and two wireless networks (one terrestrial, the other satellite), each with characteristics that differed from the ARPANET. Before long the most prevalent networks in the system were local area networks called Ethernets.

These networks and computers and, of course, packet-switching technology all formed the underpinnings of the Internet. However, the Internet would not have become a worldwide phenomenon had it not been for three critical developments along the way. First, the National Science Foundation (NSF) in the United States embraced the DARPA technology and expanded the nascent Internet so as to allow the entire research and educational community in the United States and, in due course, much of the international community to participate. Second, a competitive telecommunications industry in the United States offered several cost-effective industry proposals for a high-speed alternative to the ARPANET. Finally, the U.S. Congress passed legislation in early 1993 that permitted NSF to open the NSFNET for commercial use as well as research and education purposes.

The Internet has clearly changed the lives of us all. Worldwide communication is now cost effective and continuously available from home and office, with access to information of all kinds readily at hand. However, a number of nontechnological issues remain. Today we are concerned with matters of privacy and security as well as with intrusions such as viruses, spam, or other offensive material. There is also the question of how best to deal with intellectual property on the Internet. These are thorny problems. Still, I believe the greatest opportunity for innovation lies in the power of the Internet to inspire new forms of creativity and to foster collaboration between individuals and organizations on a far larger scale than most of us can yet imagine.