ARPANET: the night the network typed 'lo'
At 10:30 on the night of October 29, 1969, a UCLA programmer named Charley Kline sat in room 3420 of Boelter Hall and typed the letter L. Three hundred and forty miles up the California coast, at Stanford Research Institute in Menlo Park, a programmer named Bill Duvall watched it appear on his terminal. Kline typed O. Duvall confirmed it. Then Kline typed G — and SRI’s computer crashed.
The first message ever sent over ARPANET was “lo.” Not by design: just the first two letters of “login” before the remote system fell over.
ARPANET — the Advanced Research Projects Agency Network — had grown out of a 1963 memo in which J.C.R. Licklider, a psychologist turned computer scientist at MIT, imagined an “Intergalactic Computer Network” linking machines across geography so researchers could share computing power without mailing punch cards to each other. Bob Taylor, who ran ARPA’s computer science office from a single room in the Pentagon, turned that vision into a budget line: a million dollars in 1966 to build the thing. Lawrence Roberts came in to run the project. The engineering contract went to Bolt, Beranek and Newman (BBN), a consulting firm in Cambridge, Massachusetts, which built the Interface Message Processors — squat grey minicomputers that would act as switching nodes — in under a year.
The central intellectual bet was packet switching. Instead of tying up a dedicated telephone circuit for the duration of a transmission, packet switching slices data into small labeled chunks, routes each independently through whatever network path happens to be open, and reassembles them at the far end. Paul Baran at RAND and Donald Davies at Britain’s National Physical Laboratory had both developed the theory independently in the early 1960s. ARPANET was the first large-scale proof that it worked.
By the end of 1969, four nodes were live: UCLA, SRI, UC Santa Barbara, and the University of Utah. After Kline and Duvall spent an hour troubleshooting by telephone, Kline logged in successfully on the second attempt. By December, all four machines were talking.
Nobody had planned a ceremony for any of this. Leonard Kleinrock’s team at UCLA was the Network Measurement Center — their mandate was to instrument the network and record how it performed, not to announce a new era. The idea that a Department of Defense research experiment would one day carry the world’s commerce, correspondence, and arguments would have struck most of the people in Boelter Hall that night as implausible.
The killer application arrived faster than anyone predicted. In 1971, BBN programmer Ray Tomlinson sent the first network email — choosing the @ sign to separate username from host machine, a typographical decision he later admitted he gave almost no thought to — and by 1973 email had become the dominant traffic on the network. ARPANET had turned into a communications medium before anyone had consciously designed it as one.
The network was decommissioned in 1990. By then, TCP/IP — the protocol Vint Cerf and Bob Kahn had formalized in 1983 as the lingua franca for connecting different networks — had absorbed its principles into something far larger. When ARPANET shut down, Cerf wrote a brief verse in memoriam: “It was the first, and being first, was best.”
The two letters Charley Kline managed to transmit before the crash turned out to be an accidental epithet for the whole project. Lo: behold. Something was coming.
Sources
- ARPANET — Wikipedia — network architecture, the four original nodes, Licklider’s “Intergalactic Computer Network,” Ray Tomlinson’s email, the TCP/IP transition.
- October 29, 1969: Happy 40th Birthday to a Radical Idea — Computer History Museum — Charley Kline, Bill Duvall, the first transmission from Boelter Hall, Kleinrock’s account.