The Login That Broke a Computer
On October 29, 1969, at 10:30 PM Pacific time, a UCLA programmer named Charley Kline tried to do something nobody had ever done: log into a computer hundreds of miles away, over a brand-new network called ARPANET.
He typed the command "login." Two letters made it through — L and O — and the receiving machine at Stanford Research Institute crashed. Not exactly the historic moment anyone had hoped for.
An hour later, after engineers adjusted the machine, Kline tried again. This time it worked. So the very first message ever sent across what would become the internet's ancestor was, technically, just "lo." Nobody planned it. It just happened to be the last thing that survived before the crash.
The Boxes That Made the Network Real
ARPANET wasn't just wires and imagination. It ran on real machines called Interface Message Processors, or IMPs, and they were beasts even by the standards of 1969.
Each IMP started life as a ruggedized Honeywell DDP-516 minicomputer, stuffed with a mere 24 kilobytes of memory. Today that's less than a single low-resolution photo. Back then, it was enough to route data for an entire research network.
Every IMP could plug in up to four local host computers and talk to six other IMPs over leased telephone lines. By 1971, a lighter Honeywell 316 took over the job, wired to work as either an IMP or a Terminal Interface Processor, letting people type in directly from a keyboard terminal.
Four Computers, One Experiment
By December 5, 1969, ARPANET connected exactly four computers, and each one was a completely different beast. UCLA ran an SDS Sigma 7. Stanford Research Institute ran an SDS 940. UC Santa Barbara had an IBM 360/75. The University of Utah used a DEC PDP-10.
None of these machines were designed to talk to each other. That was the point. ARPANET's whole premise was proving that wildly different computers, built by different companies for different purposes, could still share a single network and pass data back and forth.
It worked. Four machines, four manufacturers, one shared line — and a template for every network that came after.
A Network That Refused to Stop Growing
ARPANET started small and never slowed down. In June 1970 it had 9 IMPs online. By September 1971 that number hit 18, connecting 23 university and government hosts. Two years later, in September 1973, it had reached 40.
The growth kept compounding. By 1981, ARPANET had swelled to 213 host computers, with a new one joining roughly every twenty days. What began as a small defense research experiment had turned into a fast-expanding web of universities, labs, and government sites, each one adding more traffic.
That relentless growth is exactly why ARPANET needed new rules for joining separate networks together — rules that eventually became TCP/IP.
The Side Project That Took Over
ARPANET was built for sharing computing power between research labs, not for chatting. But in 1971, BBN engineer Ray Tomlinson sent the first network email, and nobody could put that genie back in the bottle.
By 1973, an ARPA study found something the network's designers hadn't planned for: three-quarters of all ARPANET traffic was email. A feature nobody had officially commissioned had quietly become the network's most popular use, dwarfing the resource-sharing it was originally built for.
It's a pattern that would repeat again and again in computing history — the side feature outgrowing the original mission. ARPANET just happened to be where it started.
Turning Off the Original
ARPANET didn't disappear all at once. Through the 1980s, its role was gradually absorbed by newer networks, including NSFNET, built with support from the National Science Foundation. TCP/IP, the addressing system Bob Kahn and Vint Cerf had developed, went into full production use across ARPANET on January 1, 1983, effectively rewiring the network's plumbing while it kept running.
The project was formally decommissioned in 1990, its job done. Some of the original Interface Message Processors kept humming along until as late as July 1990 before finally being switched off for good.
By then, the private sector and telecommunications industry had already picked up where ARPANET left off, expanding it into the commercial, worldwide internet used today.
