People of ACM - Robert Metcalfe

January 16, 2024

In 1973, while working at Xerox Palo Alto Research Center (PARC), you circulated a memo with an idea for connecting some of the first personal computers—PARC’s Altos—within a building. What was the key challenge you were trying to overcome, and how did your early design for Ethernet overcome that challenge?

Dave Boggs (RIP) and I were the first to network a building full of personal computers, one on every desk. There were many constraints, but the most challenging was that our design had to fit on a board of 60 chips. These were 1973 MSI chips out of the TI 7400 series catalog. That first Ethernet ran at 2.94Mbps because we didn't have room for a clock on our board, so we used the system clock on the Alto's backplane, which ticked our Manchester encoder every 170ns.

Over the years, competing devices such as IBM Token Ring vied with Ethernet to be the main technology connecting computers. Why has Ethernet prevailed?

Ethernet prevailed because it worked, and because it was sooner, faster, and cheaper, but mostly because it was sincerely an open multi-vendor standard. My Internet startup, 3Com Corporation, shipped IBM Token Ring before IBM, but there was trouble selling it because of difficulties in installing our cards in IBM installations, which were not sincerely standard.

In recent interviews, you have contended that “the most important new fact about the human condition is that we are now suddenly connected.” In your own view, what is an important aspect of our new “connectivity” that hasn’t received enough attention?

COVID hit in 2019, the year of the Internet's 50th anniversary. It was almost as if we had built the Internet for COVID. Zoom was there and soon everyone was using it. Now many of us have had a taste of remote work and remote education. We are experimenting with the economies of massive substitution of communication for transportation. This will play out over the next decade. And the impacts will be massive and beneficial. Therefore, I like to say that COVID is short for collaborative video.

Will you tell us about the goal of your recent research into accessing more geothermal energy?

99% of the Earth's mass is above 1,000 degrees Celsius. Beneath our feet we have enough heat to power the human race for billions of years, long after Earth is scheduled to explode. So, we don't have to burn anything to make more heat. We have plenty. We just need to bring it to the surface economically. Not just clean, but cheap and clean. And so, the oil and gas industry, which knows how to drill—especially since the shale revolution—is starting what we call a “pivot” from harvesting oil and gas to harvesting heat. And they are being aided in this pivot by exciting startups. There are several kinds of geothermal sources of energy. My favorite at present is deep (kilometers) and closed-loop.

What piece of advice have you gleaned from your years of experience as an inventor that you wish your younger self knew? Or, alternatively, looking back, what was an approach to innovation that your younger self got right?

Learn how to sell. And here's the secret: listening.


Robert Melancton Metcalfe is an Emeritus Professor of Electrical and Computer Engineering (ECE) at The University of Texas at Austin and a Research Affiliate in Computational Engineering at the Massachusetts Institute of Technology (MIT) Computer Science & Artificial Intelligence Laboratory (CSAIL).

He is the recipient of the 2022 ACM A.M. Turing Award for the invention, standardization, and commercialization of Ethernet. In their classic 1976 Communications of the ACM article, “Ethernet: Distributed Packet Switching for Local Computer Networks.” Metcalfe and David Boggs described the design of Ethernet. Metcalfe then went on to guide the development of Ethernet over the next several decades. Today, Ethernet is the main conduit of network communications around the world.

Metcalfe will deliver a laureate talk at the ACM India Council's inaugural Pingala Interactions in Computing (PIC) event being held February 3–6, 2023 at the Infosys Mysore campus in India. PIC will comprise about 50 young Indian and international researchers who will interact with eminent laureates in computing.