|Claude E. Shannon||Paul Baran|
On 30 April 2016, the world will be celebrating the 100th anniversary of the birth of Claude Elwood Shannon, founder of information theory. There will be articles in The Institute and IEEE Spectrum and a double IEEE Milestone celebration of Shannon’s 1948 paper demonstrating the nature of digital communication and the possibility of error-free transmission (at Bell Labs in New Jersey and MIT in Massachusetts), among many other recognitions around the world.
So, rather than adding to that chorus, I thought I would remind readers that, despite the clear groundbreaking importance of Shannon’s work, it took the effort of many others building on it to produce the Internet Age in which we live. By an interesting coincidence, just one day before the Shannon centennial, 29 April 2016 will be the 90th anniversary of the birth of one such digital communication pioneer, Paul Baran, inventor of packet switching. Note that packet switching was slightly later invented independently by Donald Davies, who coined the term—Baran had given it the unfortunate name Distributed Adaptive Message Block Switching—but it was Baran who first described the concept. Packet switching is the method of breaking a digital message into blocks, sending them by multiple simultaneous channels, and reassembling them, thus increasing network robustness and efficiency. This is the basic concept upon which the Internet operates.
Baran’s birth, almost ten years to the day after that of Claude Shannon, could not have been under more different circumstances. Shannon was born in Michigan to a successful businessman, a descendant of New England Puritans and a distant cousin of Thomas Edison. His mother was an educator. Paul Baran was born in Grodno, Poland (now in Belarus), to a poor Jewish family. When he was two years old, the family immigrated to the United States and ended up in Philadelphia, where his father opened a grocery store and his mother was a housewife. Shannon attended the University of Michigan, graduating with two bachelor’s degrees in 1936, and then M.I.T., where he obtained his Ph.D. in 1940. Baran went to Drexel Institute of Technology (now Drexel University), receiving his B.S. degree in 1949, and after working for a while he then pursued graduate work at UCLA under Gerry Estrin. However, his need to continue to work full time led him to withdraw with only a Master’s degree, in 1959. At that point, he joined the RAND Corporation, and, this being the height of the Cold War, was put on a team working on the problem of designing a communications system that could continue to operate during a nuclear exchange.
Obviously, the concept of packet switching of digital signals builds on Shannon’s overall theory of communication but, as Baran points out in his IEEE oral history, it was another aspect of Shannon’s work that caught his attention. Shannon was famously interested in a wide range of engineering issues beyond communication theory. In 1950 at Bell Labs, he produced an electromechanical mouse that could learn its way through a maze, and then apply its previous knowledge when the maze was reconfigured. This feat is considered by many to be a forerunner of artificial intelligence.
Baran was thinking about solving the robust network problem by breaking up messages and having the pieces “route” themselves when he read of Shannon’s mouse and realized it had relevance to his own work. What if the mouse was a message trying to get from one end of the maze—the communication network—to the other? The information, like the mouse, could learn as it moved through the system. But to have any efficiency it needed to know not just how to remember, but how to forget when a particular part of the maze was reconfigured—that is, destroyed by a nuclear blast. Baran now had all the pieces he needed, and, in 1964, published On Distributed Communications. In 1968, ARPA issued a “request for quotation” for a distributed computer network. BBN got the contract, and worked with Leonard Kleinrock (his earlier work on queuing theory was critical in making packet switching truly work) of UCLA, and the ARPANET went live in 1969. The Internet Age was dawning.
Both Shannon and Baran continued to work past traditional retirement ages, although Alzheimer’s disease limited Shannon’s career, and he died on 24 February 2001. Baran stayed active much longer. His contributions included work on early data encryption standards, developing the first commercial asynchronous transfer mode system, and founding one of the first wireless internet access services. Paul Baran continued to work right up to his death on 26 March 2011, one month shy of his 85th birthday. So we have a lot of engineering history to commemorate this April.