The 2022 Turing Award is awarded to the 76-year-old father of Ethereum! A failed doctoral thesis made the Internet era a reality

Just last night, ACM announced the winner of the 2022 Turing Award - Ethernet inventor Bob Metcalfe, in recognition of his invention, standardization and commercialization of Ethernet.

The ACM AM Turing Award is known as the "Nobel Prize in computing", and the winner will receive a reward from Google $1 million prize.

The significance of the invention of Ethernet to mankind is obviously self-evident. This technology, developed 50 years ago, is the foundation of the Internet, allowing our computers and billions of other devices to connect to the Internet, and making them interconnected.

Jeff Dean, vice president of artificial intelligence at Google, said in the official ACM announcement, "Today, there are approximately 7 billion ports in the world, Ethernet is everywhere, and we all take it for granted. .However, it is easy to forget that without Bob Metcalfe’s invention and his enduring vision that every computer must be connected to the Internet, our connected world would be very different.”

Yes, before the invention of Ethernet, the world was not flat. At that time, people never imagined that 50 years later, a robot named ChatGPT would conquer the world through this network.

A failed thesis

This invention that has affected all mankind originated from a failed graduate student thesis.

Metcalfe, born in 1946, studied electrical engineering and industrial management at MIT as an undergraduate.

After graduation, Metcalfe went to Harvard for graduate school and was also hired as a researcher at MIT. This coincided with the U.S. Department of Defense increasing its focus on Arpanet (a precursor to the Internet). invest.

He proposed to Harvard that an interface should be established to connect Arpanet to Harvard's mainframe computers, but Harvard rejected him. He turned to MIT and made the same suggestion.

Metcalfe, while still a graduate student, built this interface to connect MIT's mainframe computers to the Arpanet, the precursor to the modern Internet

In 1972, he submitted his thesis to Harvard's thesis committee describing his idea, but it failed to pass the defense. The committee felt that the topic was not theoretical enough.

Standing on the shoulders of those who came before us

Fortunately, geniuses always have a place to display their talents.

By then, Metcalfe had been offered a job at Xerox Corporation's Palo Alto Research Center (PARC) in California.

Lab director Bob Taylor told him that he had to come to Palo Alto to complete his thesis no matter what.

Once there, Metcalfe began building another Arpanet interface for a new PARC computer, while trying to find a topic "theoretical" enough to satisfy the Harvard thesis committee.

Xerox’s PARC experiment hoped to create the first modern personal computer and have one on every desk.

Metcalfe said, "So they asked me to design a network for this. I got a network card with 60 chips. That was the second best thing in my life. Things. The first one was for me to be born into my parents' family."

MetCalfe in Texas Austin's home, holding his 1973 doctoral thesis

#At that time, computer networks were both an engineering challenge and a theoretical challenge. The fundamental difficulty lies in sharing access to the network among many users.

The telephone network solves this problem in the simplest way: during a call, the connection between the two parties locks the communication channel so that other users cannot access the channel, even if the channel is not fully utilized.

However, this inefficiency is not a big problem for the telephone network. When you make a phone call, there is always one person talking and the call will not fall into long silence.

But computers are different, their communication relies on short pulses, and these pulses are usually separated by a long dead time.

How to describe the data flow in the network?

In the early 1960s, computer scientist Leonard Kleinrock showed that this problem could be solved using queuing theory. The theory is a branch of mathematics that models traffic jams and other things that can happen when people wait in line.

Paper address: https://dspace.mit.edu/handle/1721.1/11562

This model showed engineers how to significantly reduce dead time, and Arpanet was able to prove that this approach actually works in practice. However, how to coordinate traffic in the network is not an easy task.

In 1971, Norm Abramson, a professor at the University of Hawaii, demonstrated a radical alternative to transportation coordination that would make any City planners were shocked.

He built a radio network called ALOHAnet, which, like Arpanet, transmitted data in tiny packets.

But unlike Arpanet, ALOHAnet does not try to avoid collisions between packets. Instead, any user whose messages are lost or garbled due to a conflict can simply try again after a random interval.

This "random retransmission" is similar to the etiquette in people's conversations: when two people start talking at the same time, both parties stop and try again later. After a few attempts, the problem will resolve itself.

However, this strategy also has a bug: under low traffic conditions, the transmission is smooth, but when the network becomes congested, collisions become so frequent that there is no No more messages can come through.

Metcalfe stumbled upon a paper by Abramson that devised a way to bypass the ALOHAnet queuing theory deadlock.

Paper address: https://www.clear.rice.edu/comp551/papers/Abramson-Aloha.pdf

Inspired by this, Metcalfe built a model in which users independently adjust the average waiting time between transmission attempts, while also taking into account the frequency of conflicts. If the number of collisions is rare, they will retry faster; if the network is congested, they will drop out to maintain overall communication efficiency.

Abramson’s paper illustration

This The model finally made Metcalfe's paper "theoretical" enough for him to pass his defense at Harvard University. At the same time, Metcalfe also realized that he could put this model into practice at work.

At the time, Arpanet allowed researchers to share large computers, but it was expensive. ALOHAnet also connects many access points to a central hub. And the PARC laboratory is pursuing a special approach to computer networking.

Abramson’s paper illustration

PARC Taylor had an idea: a local network composed of many computers in the same building. As a new employee, Metcalfe immediately began to design this network.

Ethernet, which will forever be remembered in history, was born

In a famous memo in May 1973, Metcalfe elaborated on his vision.

On the first page of the May 1973 memo, Metcalfe laid out his vision for Ethernet

He proposed that the first personal computer - PARC's Altos - be connected in a building, which would pave the way for devices to communicate with each other and share information in a local area network. Leveled the road.

The first Ethernet ran at 2.94 megabits per second, approximately 10,000 times faster than the original terminal network.

The memorandum recommends that the network should adapt to new technologies such as optical fiber, twisted pair, Wi-Fi, and power networks, and replace the original coaxial cable as the main communication method with "Ethernet" network".

This contribution was later immortalized in Metcalfe and colleagues' 1976 ACM Communications article "Ethernet: Distributed Packet Switching for Local Computer Networks."

Paper address: https://dl.acm.org/doi/pdf/10.1145/360248.360253

On the left is Bob Metcalfe; on the right is David Boggs

In the memo, his idea combined Abramson's random retransmission system, Metcalfe's adjustments to timing, and other improvements to the ALOHAnet model that mitigated the effects of collisions.

Some of these theoretical innovations came from other researchers, but Metcalfe was the first to integrate them into actual local network designs.

In his idea, the central hub of ALOHAnet was also eliminated. Instead, the computers will be connected through some passive medium.

Metcalfe thought of a specific cable that was very practical. But he noted that, in theory, other wired or wireless networks could work just as well.

To avoid emphasizing specific hardware, Metcalfe called his idea Ethernet, which is now known as Ethernet.

"Ether" is a medium that physicists in the 19th century hypothesized electromagnetic waves would pass through. This idea was later proven wrong. But the word has endured throughout history.

One of Metcalfe’s prototype Ethernet adapters that connected early personal computers to the computer he helped create in 1973 Developed on the first local network

In November 1973, Metcalfe and his colleagues built the first network.

He continued to further develop the network, hoping to expand it beyond The pace of technology commercialization is extremely slow.

Forever Young, Always Trying

By 1979, Metcalfe had finally had enough. He left PARC and founded his own company, 3Com, to do what he couldn't do at Xerox.

Vinton Cerf, known as the father of the Internet, described Metcalfe this way - "You would never use humble to describe him. Once he had an idea, he would immediately Put it into practice.”

Soon after starting his business, Metcalfe shook up the entire industry. It convinced original companies Xerox, Intel, and many now-defunct digital equipment companies to adopt Ethernet as an open industry standard for local networks.

In the competition with other technologies, Ethernet eventually won. It can be said that this is due to its suggestibility and the standardization promoted by Metcalfe in its early days.

3Com greatly increased the commercial viability of Ethernet by selling network software, Ethernet transceivers, and Ethernet cards for small computers and workstations. When IBM introduced the personal computer, 3Com introduced the first Ethernet interfaces for the IBM PC and its clones.

In 1990, Metcalfe left 3Com and became a pundit and technology columnist. While continuing to work as a venture capitalist, he later served as a professor at the University of Texas at Austin.

Why is there such a big change in career? Metcalf has a theory about this: "In the beginning, you don't know anything, and then you go up the learning curve, and then you know everything." And he pointed out that the most interesting part is in the middle of the curve.

Over the years, Ethernet has undergone many adjustments, with few of the original technical details retained. But it continues to play an integral role as the home conduit for today's personal computer networks.

It’s Ethernet that makes everything possible.

Now, the 76-year-old Metcalfe still hasn’t stopped tossing. He changed career tracks again last year and is now a researcher at MIT studying supercomputer applications in energy and other fields.

The 23-year Turing Award winner said, "I'm still in the early stages of the learning curve. I don't know much, but I'm still working hard."

Today, Ethernet is the primary channel for wired network communications worldwide, handling data rates from 10 Mbps to 400 Gbps, with the emergence of 800 Gbps and 1.6 Tbps technologies.

Ethernet has also become a huge market, with revenue from Ethernet switches alone exceeding $30 billion in 2021, according to International Data Corporation.

And because of old times, Metcalfe insists on using the original name "Wireless Ethernet" to call Wi-Fi.

Professor Daniela Rus, director of CSAIL and professor of MIT’s Department of Electrical Engineering and Computer Science, said, “Because of the invention of Ethernet, he achieved seamless communication and information sharing, which has enabled countless people to become our It has paved the way for applications that are integral to our daily lives. From the Internet to online video streaming, Ethernet forms the basis of modern technology and has transformed the way we connect. Without the connectivity that Ethernet brings, it would be difficult to understand Life."

And Bob Metcalfe himself said: "Now, all human beings are suddenly connected together. This is the most important fact about the condition of all mankind."

The above is the detailed content of The 2022 Turing Award is awarded to the 76-year-old father of Ethereum! A failed doctoral thesis made the Internet era a reality. For more information, please follow other related articles on the PHP Chinese website!