Thirty years ago this month, Bob Metcalfe first proposed the idea of Ethernet at Xerox Palo Alto Research Center (PARC). Ethernet's acceptance as a fundamental standard seems almost foreordained today, but it wasn't always so. Metcalfe went on to found 3Com and was CEO of InfoWorld Media Group as well as a longtime columnist for this publication. Today he is a general partner in the Boston office of the IT venture-capital company Polaris Venture Partners. He spoke with InfoWorld Test Center Senior Analyst Wayne Rash about how Ethernet has surpassed even some of his expectations. (Participate in a live online forum with Metcalfe on Thursday, May 15th, at 10 a.m. PDT.)
If 30 years ago, when you invented Ethernet, I had said to you, one of these days Ethernet is going to be running at 10Gbps, what would you have said?
I don’t think so. The Ethernet we built starting in May of '73 ran at 2.94Mbps on coaxial cable. We thought that was overkill. It would go faster than anyone would probably need. Then in 1976, we built a 100Mbps optical-fiber Ethernet hub.
Why do you think Ethernet prevailed and other standards fell by the wayside?
The reason that Ethernet has endured is it’s based on a unique set of ideas, which have a lot to do with the technology and more to do with the business model. That is, Ethernet is based on de jure standards not de facto standards. And unlike Linux, it’s based on the technology and the competitors owning their implementations rather than giving them away. It’s based on fierce competition to implement the standard. On the other hand, the Ethernet-type marketplace is based on interoperability among products. The standard evolves with market experience, as we’ve seen Ethernet go from coaxial [to] twisted pairs [to] optical fibers, from 10Mb to 100Mb to 1Gb to 10Gb, from shared to switch. It’s a rapidly evolving standard. Finally, that evolution is disciplined to preserve as much as possible the installed base. And that combination of ideas is what has given the Ethernet the unexpected robustness that we’ve seen.
Ethernet didn’t seem to have the problems that bedeviled IBM's Token Ring and ArcNet. Did this play a role in Ethernet's success?
Yes. Ethernet was simpler than its alternatives. It didn’t, for example, rely on the circulation and maintenance of the token. And Token Ring mathematically and, I think, sometimes unrealistically had advantages over Ethernet as long as the token was humming along. But if that token ever got lost or duplicated to create a storm, then things went to hell in a hand basket pretty quickly. The simplicity also allowed the implementations of Ethernet to be simpler and cheaper and come to market sooner. I mean, Ethernet beat Token Ring to market by five years. So that’s quite a lead. That’s probably what saved Ethernet from the IBM juggernaut.
What did you think of the claims made that Ethernet would only run at 20 percent of capacity?
Well I was disenchanted. I became cynical and depressed …. It was a propaganda war …. I watched people say things about Ethernet that just were not true. And people believed them …. In any case, I sometimes think IBM fell from power, fell from grace in the mid-1980s, and I believe Ethernet was the straw that broke the camel’s back. That was the last time IBM tried to make an industry standard by sheer force of monopoly power.
If you could project the future of Ethernet, tell me what you think it would be?
The Ethernets that we’re talking about today bear little resemblance to the CSMA/CD (Carrier Sense Multiple Access/Collision Detection) co-ax shared thing that Dave Boggs and I built in 1974. Ethernet will begin playing a role in bridging the telechasm. Forget T1, which is an overpriced service offering from the copper monopolies. Speeds are going to keep coming and there will be all sorts of variations. There’ll be something else faster and probably more cellular, with the frequency reused through spatial division, that is smaller and smaller.