Electronic Design
Robert Kahn: Expect Much More Computer Networking To Come

Robert Kahn: Expect Much More Computer Networking To Come

Kahn (2009)

“We’ve barely seen the tip of the iceberg with the creation of a national information infrastructure,” said Robert Kahn, who helped create the ARPANET, the pioneering computer network. With Vinton Cerf, he also created the TCP/IP communication protocols that make the Internet possible.

“I think there are still significant challenges that face us in evolving the national infrastructure to incorporate new technical capabilities. Come back in 100 years and you probably won’t recognize the Internet,” he added.

Kahn had been intrigued with the possibility of connecting networks and computers for some time before working at the birthplace of the Internet, the Defense Advanced Research Project Agency (DARPA).

“In the 1960s, most computers were large batch processing machines. Programming was done on punched cards and the results printed out the next day. There were only a small number of timesharing computers by the sixties, and organizations that did not have one occasionally rented time on commercial services. It occurred to me that linking such interactive systems together over a computer network would be an interesting idea,” said Kahn.

With support from DARPA and a highly competent team of implementers at Bolt Beranek & Newman (BBN), where he worked at the time, that’s what he did. And the world soon learned about it.

Kahn arranged a week-long public display of the ARPANET, with experts interacting with approximately 40 distributed and networked computers at the International Computer Communication Conference (ICCC) in Washington, D.C., in October 1972.

THE BACKGROUND
Kahn obtained a BEE from the City College of New York in 1960 and an MA and PhD from Princeton University in 1962 and 1964. Prior to entering Princeton and during summers thereafter, he worked at AT&T Bell Laboratories. After graduation, he joined the faculty at the Massachusetts Institute of Technology, where he was advised to obtain some real-world experience while pursuing an academic career.

So, he took a leave of absence from MIT in 1966 and went to BBN, where he wrote a number of technical reports on the various aspects of computer network design. He decided not to return to academia at the time. While at BBN, a DARPA request for quote regarding development of a computer network appeared on his desk. It piqued his interest, and although he didn’t expect to be involved himself, he was tapped to help write the technical proposal to DARPA.

When BBN won the contract to build the ARPANET’s packet switch, or Interface Message Processor, and deploy an initial fournode network, Kahn was responsible for the system design. Then, in 1972, Larry Roberts persuaded Kahn to join him at DARPA. A week before joining DARPA, Kahn put on the networked computer display at the ICCC event.

While at DARPA, Kahn was responsible for the development of two different packet networks: a packet radio network (using embedded microcomputers) and a satellite packet network (later deployed on Intelsat IV with one ground station in the U.S. and several ground stations in Europe). These networks were in addition to the ARPANET, which used land lines leased from AT&T. The packet radio net utilized direct-sequence spread-spectrum technology, similar to the technology in today’s well-known CDMA systems—another first.

SUCCESS LEADS TO THE NEXT STAGE
The availability of these three networks presented a challenge. Now that different computers were successfully grouped together in a network, i.e., the ARPANET, why not have the three different networks work together? If they could all interoperate, users on the packet radio net, for example, could use computing resources on the ARPANET. And researchers on the ARPANET, or the packet radio net, could interact with their counterparts in Europe.

“I wondered, ‘How can these packet-switching networks be connected so that computers on any of the networks could communicate?’ That is where the idea came for the open-architecture protocols,” said Kahn.

Kahn was initially concerned about two basic issues in connecting potentially dissimilar networks. One issue concerned addressing. A means was needed to identify machines on other networks, which led to the creation of what are now known as Internet Protocol addresses, or simply IP addresses. The second issue concerned the potential loss of information en route. The existing ARPANET host protocol was not designed for loss of information or for communicating in a multiple-network environment. Clearly, a new protocol was needed.

Kahn began to develop an open-architecture approach to internetworking that allowed different kinds of packet networks to communicate with each other. Shortly thereafter, he invited Vinton Cerf, then a new assistant professor at Stanford University, to work with him on the project.

“He knew about computer operating systems and how to embed protocols in them. I knew about the communications and systems aspects. Although I had been working on the basic architectural design of the system, I didn’t know how best to handle the variety of networked computers then using the ARPANET, and Cerf knew how to do that. Together we were able to improve on the basic architectural ideas and protocols,” said Kahn.

Subsequently, as director of DARPA’s Information Processing Techniques Office (IPTO), Kahn initiated the largest computer research and development program ever undertaken by the U.S. Government, the Strategic Computing Program.

Kahn left DARPA after 13 years in late 1985. Shortly thereafter, he formed the Corporation for National Research Initiatives (CNRI), a not-for-profit organization to promote research and provide leadership and funding for network-based information technologies and information infrastructure development. In addition, through its MEMS and Nanotechnology Exchange, CNRI assists researchers in the U.S. in prototyping microelectromechanical systems (MEMS) and nanotechnology devices.

IN REVIEW
One would expect Kahn to be rightfully proud of his critical role in the Internet’s creation and his contributions to the development of computer networking, but that’s not what he’s most grateful for. Rather than cite a specific accomplishment as his major achievement, he’s particularly proud of “working effectively with the research community to help move forward a research agenda in information technology.”

“With my colleagues at DARPA, we were able to support many different areas. We were moved forward on the networking front, particularly with the development of new networks and, of course, the Internet. We supported important research in parallel computation, advanced software, artificial intelligence, and design of very large-scale integration (VLSI) devices and systems. I’m grateful that our efforts allowed the computer research community to bring such enormous talent together to enable innovations in information technology,” he said.

“We made a broad technology push across multiple fronts and achieved significant progress by leveraging the talents of the research community,” said Kahn.

TODAY’S GOALS
Kahn is now chairman, CEO, and president of CNRI. In this capacity, he strives to formulate and promote pilot projects focusing on national infrastructure issues like managing information in an Internet environment over time frames that may range from seconds to millennia. “I spend most of my personal research time in this general area,” said Kahn.

“With support from the Department of Defense, CNRI has been helping to stimulate research in the MEMS and nanotechnology area, where there’s lots of ideas and various existing and potential commercial uses,” said Kahn. Also at CNRI, Kahn and Cerf coinvented Knowbot programs, which are mobile software agents in the network environment.

PROMOTING RESEARCH FOR TOMORROW
Digital objects and the Handle System also capture a large part of CNRI’s focus. The digital object architecture provides an open-architecture approach to managing digital information in an Internet environment. Digital object identifiers, also called Handles, refer to unique persistent identifiers for digital objects. The Handle System manages such identifiers and provides useful metadata about the digital objects they identify.

This would be helpful, for example, with digitized archives or the “electronic medical records” favored by the Obama administration. Electronic medical records, Kahn said, would be valuable if you became ill overseas and your medical history could not otherwise be made available quickly, for example. Or, they would be valuable when dealing with multiple medical facilities and multiple doctors. Duplicate tests, for instance, might be avoided.

“A major expenditure by the private sector will also be needed to take advantage of electronic medical records, even if the infrastructure were in place. We shouldn’t assume the federal government will be paying 100% of the cost,” said Kahn.

ADVICE TO STUDENTS
“Do the best you can, understand what motivates and excites you, and trust your own instincts,” Kahn tells today’s students. Lots of people, Kahn said, advised him that going into networking was not a smart idea. Yet he persisted because he was interested in it and trusted his own instincts that the research area would ultimately prove beneficial.

Hide comments

Comments

  • Allowed HTML tags: <em> <strong> <blockquote> <br> <p>

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.
Publish