In 1966, when Bob Taylor and Harp planted the seed of something big.
They planned to connect 3 mainframe computers that were far apart. One was at the Massachusetts Institute of Technology (MIT), more than 700 km away. away
The other two were on the other side of the country. One at the University of California and the other at the Strategic Air Command in Santa Monica, or Q32.
Each of these massive computers required a login procedure and a different programming language.
Next to his office was the terminal room. A small space in which there were three remote access terminals, side by side.
Although Taylor could access these computers remotely through their terminals. They could not easily connect with each other, or with other Arpa-funded computers in the United States.
Sharing data, dividing a complex calculation or even sending a message between those computers was almost impossible.
A common language
The physicist Wesley Clark came up with a solution, using a common language. Clark had been following the emergence of a new generation of computers.
The minicomputer was modest and economical. Not like the mainframes of the size of a room installed in universities in the United States.
He suggested installing a minicomputer in each site of this new network. The local mainframe, the huge Q-32, for example, would communicate with the minicomputer that was nearby.
The minicomputer would be in charge of communicating with all other minicomputers in the network and would be responsible for the new and interesting problem of moving data packets reliably through the network until they reached their destination.
All minicomputers would work the same way, so if a network program was written for one, it would work on everyone.
Existing mainframes would continue to do what they already did well.
The new minicomputers would be optimized to reliably manage the network without fail.
Clark’s idea was that each local central unit had to be programmed simply to communicate with the “little black box” next to it: the local minicomputer.
That was the only thing it would take for it to be connected to the entire network.
The “little black boxes” were actually big and gray, called Interface Message Processors (IMP).
The IMPs were customized versions of Honeywell minicomputers, with the size of refrigerators and weighed more than 400 kg. each.
What the designers of the network wanted were message processors who worked in silence. With minimal supervision, and did not stop despite the heat or cold, vibrations or surges. Honeywell military-grade computers seemed the ideal starting point, although their shielding was perhaps a bit excessive.
“Behold”
The prototype, IMP 0, was ready in early 1969. But it didn’t work. A young engineer set about fixing it for months, manually unwrapping and wrapping wires around metal sticks separated by a distance of approximately 1 millimeter.
It was not until October of that year that IMP 1 and IMP 2 were ready at the University of California, Los Angeles, and the Stanford Research Institute, more than 500 km away.
On October 29, 1969, two central computers exchanged their first word through their complementary IMPs.
The word was: “Lo” … The intention of the operator had been to write: “Login” (“Login”) but the network fell after two letters.
A rough start, but Arpanet was turn on.
Other networks followed, as was a decade-long project to interconnect them into a network of networks, or simply, “internet.”
Finally, IMPs were replaced by modern devices called routers. But the world that had been predicted, in which almost all conceivable elements of computer hardware and software will be on the network, was coming true. And the IMPs had opened and showed the way.
And harp? Thanks to the clarity of their mission, them also played an important role in the creation of the global positioning system and, more recently, driverless cars.
https://www.internetsociety.org/internet/history-internet/brief-history-internet/
