Metcalfe on inventing Ethernet with Dave Boggs in 1973

本站添加: 2026-06-03

[00:00:01]I decided that we should build bit serial bus.

[00:00:06]We would run instead of home running every terminal to this room, and we had these in the building and there were horrible rats nest of cables and I took photographs of them.

[00:00:16]I decided we would build a network that didn't have the centralized feature.

[00:00:20]This would be a completely distributed network, maximally distributed. That was the goal.

[00:00:27]Maximally.

[00:00:29]So we I started fiddling around with how to do that.

[00:00:34]I remember I took I bought I ordered up some coaxial cable, 1,000 ft of it, you know, a spool and and a big event happened while I was down there that day.

[00:00:47]Um this was coax and I wanted I needed to put connectors on the end to hook it up to the oscilloscope because I wanted I wanted it to be a high quality connection so I could see all the capacitances and the inductances and all that stuff. I wanted to see what it really looked like.

[00:01:02]So I started putting connectors on the ends of this coax and there was another guy hippy type guy with actually hard he was wearing a Brooks Brother yellow button down shirt just like I used to wear.

[00:01:17]It was David Boggs.

[00:01:19]And he was um bringing up um Nova 800s as a technician. He was a grad student at Stanford.

[00:01:25]And he saw me trying to put the connectors on the cable and he said I know how to do that, let me do it. And it involves you know, skinning it and clamping it and crimping it and soldering it and a bunch of things that he had done a lot of and I hadn't done much of. So he in that moment we became buds.

[00:01:45]So Boggs and I formed our partnership right there.

[00:01:49]So we started launching and we could see that you could send digital bits down a cable, not a breakthrough discovery, but for us we we got really practical about it.

[00:02:02]So we So on May 22nd, 1973, I wrote a memo in which Ethernet was named. It was the formally had been called we had been as a working name calling it the Aloha the Alto Aloha network.

[00:02:20]And in that memo I renamed it the Ethernet with a capital N and a space ether space net. The goal here was to have it be maximum have it the network be maximally distributed. There was these workstations these Alto this these prototypical personal computers going to be one in every desk. Can you imagine that a computer on every desk? Wow.

[00:02:45]Very controversial in 1973 that you would why would why would you want a computer on every desk when you have this time sharing system down in the basement? Why would you want a computer and your wife was what possible use could there be for such a thing? I remember people had that discussion.

[00:03:00]But I I was bought into maximally distributed. So I wanted as little in the middle of the network as possible and everything on the edges.

[00:03:07]So the middle was going to be the medium was going to be this cable and we settled on coaxial cable cuz it was there was a pre-existing device the vampire tap it was called that would allow you to puncture it without cutting it.

[00:03:21]And but we could have used twisted pair and we could have used optical fibers we could have or radio.

[00:03:27]I decided that I did that Ethernet would work on any medium.

[00:03:33]Cuz all we required was that the medium be passive omnipresent and a medium for the propagation of electromagnetic waves. That's all it needed to be and it could be coax it could be twisted pair. So instead of calling it coax net we had to I wanted to think of something more abstract than that.

[00:03:57]And I remembered from freshman physics at MIT about the Michelson and Morley experiment in which these two scientists performed this very famous experiment in which they proved that the luminiferous ether did not exist.

[00:04:14]And what was the luminiferous ether? It was omnipresent completely passive and it served as a medium for the propagation of electromagnetic waves. Da-da.

[00:04:25]Ether net.