Bob Metcalfe

May 21, 2013

You’re probably connecting to reddit through a technology I invented. I’m Bob Metcalfe and I invented Ethernet – AMA

On May 22, 1973 with David R. Boggs, I used my IBM Selectric with its Orator ball to type up a memo to my bosses at the Xerox Palo Alto Research Center (PARC), outlining our idea for this little invention called “Ethernet”, which we later patented.

I worked with the IEEE Standards Association to develop the IEEE 802.3 standard for Ethernet, which specifies the physical and lower software layers. Today Ethernet and the IEEE 802.3 standard are the foundation for today’s world of high-speed communications used in billions of homes and businesses around the world.

I submitted this to the mods awhile back so I could get on the calendar but I figured you’d like to see it, too. Now, ask me anything!

It's been two hours and 179 comments. Have to go now. For more about Ethernet's 40th Birthday, go to http://www.facebook.com/Ethernet40thAnniversaryIEEESA



How do you feel about Software Defined Networking? Do you think we will see a trend of moving to centralised control of traditionally distributed algorithms in networking?

SDN is one of the next big things in the Gigafication of the Internet. Control is moving into the network, but I would not say it is being centralized.


Do you get royalties on your invention?

I made a play to get a buck per Ethernet node, but had to settle for a penny per packet.


He founded 3Com. If you do a Google search of him it says he made about $250 million.

Do not believe everything you read on the Internet.


I like to imagine you go out to drink with David Reed, bickering who's laws have the greatest validity and contribution to different forms of networks. Please confirm or deny.

Seriously, though, what is the most exciting way you have seen somebody embellish your own work?

Have not seen Dave Reed in years, but am a big fan of his law, which is even more of a gross exaggeration of the Network Effect than mine. Metcalfe's Law needs refinement, but let's do it with data this time.


What is the best/biggest thing you've done or witnessed as a Member of the MIT Corporation?

Attending MIT trustee meetings is the most fun you can have standing up. Recent excitement was the debut of the MOOC edX, which is going to help the Internet disrupt education the way that iTunes disrupted music and Amazon disrupted books, or BOOCs as I call them.


What do you think the most positive impact the internet has had on society as a whole?

The Internet reduces market frictions and expands freedom of choice. I give the Internet credit for everything good that has happened since 1969.


Do you think the claims that the internet reduces actual communication and devalues some of the things that used to define society are valid?

No. But I think the there's good stuff on TV, more good stuff than before, despite all the crap. Good thing we have search.


You invented a product which is used extensively around the world and yet many people (me included) had no idea who you were (sorry). Do you find this annoying?

Am quite famous among my people, networking nerds. That's enough for me. On the other hand, who is Katy Perry?


Do you think that UT-Austin will ever be a top 10 university overall in the U.S.?

UTAustin is already top 10, depending on how you count. Hook 'e, Horns!


What is the craziest thing or product that uses Ethernet?

Everything uses Ethernet, so it's hard to pick one.


Everything uses Ethernet, so it's hard to pick one.

F-16.


How do you like Reddit? How do you feel about what you've done? (Making this great invention)

edit: I'm special

Jury is still out on Reddit. And if it's Ethernet you're asking about, I feel happy, grateful, proud, and wary about what you are going to say next. The narwhal bacons at midnight.


Xerox PARC made a ton of innovations that shaped the modern world. So did Bell Labs, and DuPont Labs and other corporate research labs. Many of these labs have closed or shrunk drastically. Which corporate research divisions are shaping the future now?

I think the future of research will be at research universities supported by government agencies, especially NSF. Universities graduate students, who have proven the most effective innovation vehicles.


Would you rather fight one horse sized duck, or 100 duck sized horses?

The bigger they are the harder they fall.


What do you think separates intellectual innovators from the rest of the population, for example, Mark with Facebook? It seems like such a simple website that anyone could've started. What prompts the mind to such great ideas , or new innovations? In other words, what advice would you have for an undergrad that wants to make things that literally change the world?

Innovations depend much on context, and so it helps to be at the right place at the right time, as Zuckerberg is. But then you have to be skilled enough and ambitious enough to act, as Zuckerberg has.


In movies, hackers often clip a device to Ethernet cable to steal or inject information through the insulation and cable shielding. This is impossible, right?

Thanks for the great invention!

No, not impossible, just difficult. Security should not be implemented at the hardware level. Higher-level protocols should be relied upon, not cable insulation.


What do you think will be the successor to ethernet?

For decades now, when a new networking technology proves out, they call it Ethernet, except for WiFi, which started life as "wireless phy Ethernet." The PARC CSMA/CD coaxial cable Ethernet has already had many successors.


Thank you for doing this AMA!

What are your thoughts on the future of Internet privacy and government control of information?

Trust governments to invade your privacy. We must use tools to keep our stuff a secret. Am not expert on this, but I do mail all my financial secrets to the IRS through the USPS every April 15th trusting that no USPS union member or IRS agent will peek. Oy.


What are your thoughts on Google Fiber (as it's coming to Austin soon)? And what, exactly, sparked your interest in communications?

Also, I'd like to thank you for your speech at my graduation (from UT) last year, it was very inspiring. When I return for grad school I hope to have a chance to work with you.

Google Fiber is great news for everyone, especially as a spur to AT&T and Comcast and Time Warner et al. Competition! We are now gigafying the Internet -- build it and they (new apps) will come, so far anyway.

Got interested in communications because that's what ARPA was funding the year I started grad school in 1969.


Hi Bob. I was one of the founding engineers at Archivas, which you decided to fund, when you were a VC in the Boston area. We gave an overview of the technology, and while it apparently went well, you noted the low-key nature of the group, and called me out in particular for being "mopey".

No question here, I just wanted to thank you, both for investing in Archivas, and for giving me a great anecdote, that has now become both a family legend, and a mark of distinction for me in my circle of acquaintances in the local industry.

The father of Ethernet called me mopey!

Not anymore. How about snarky?

Thanks for making us so much money with Archivas.


Has the Internet exceeded your expectations when you first created it, and if they have, what were your expectations? Did you think that it would reach the level that it has today?

By far, more each year, who would have guessed? We were building our own tools, and they escaped to serve uses unimagined, say like YouTube.


Hi Bob, can you to speak to new markets that Ethernet will be expanding to?

Ethernet is going up, into, over, across, and down into new markets. Up toward terabit LAN. Into the WAN killing SONET. Over the airwaves as WiFi. Across the telechasm, between carrier WANs and customer LANs, as Carrier Ethernet. And won into embedded networking, as ZigBee (IEEE 802.15.4).


Did you make a lot of money off the invention?

Did not make my Ethernet money on patent royalties, but by SELLING Ethernet for a decade to people who didn't know they needed it.


How badass does it make you feel to be able to honestly say "I invented Ethernet"?

You are truly a pioneer, you must be proud.

For some values of I, invented, and Ethernet, I can honestly say that I invented Ethernet. But so can a lot of other people. Proud, yes. Also, wildly curious about where this monster goes next.


if you are confident you won't have to eat your words, any new predictions?

Better to eat my words than someone else's. I make 10 predictions per day. My batting average is above .500. Y'all of course remember the big ones I got wrong.


How do you deal with challenging people or situations? What is the best advice someone told you?

God (or Darwin) gave us one mouth and two ears. Take the hint. Best to listen first. Summarize back with the language you've heard. Then, act!


I knew this picture would finally come in handy for lots of karma! But seriously, thank you for all you've done for the world, for technology, and especially what you've done for me personally at UT. I meant to ask you this but I never got the chance, out of all of the careers you've had so far, which has been your favorite and most rewarding?

Five careers, not counting 23 years as a student: engineer-scientist, entrepreneur-executive (when my company grew too big), publisher-pundit, venture capitalist, not professor of innovation. Favorite? All so different, they defy comparison. Just another 7.5 years to next career.


there's a lot of competition around messaging now. who do you think will have the biggest share two years from now? (facebook, google, whatsapp, apple)

'Messaging?" Facebook (because my daughter works there.)


High, im sitting in class browsing reddit thanks to you, how do you feel your invention has impacted the younger generation?

Close your PC and pay attention to the professor. And do not get me started on ageism with this "younger generation" stuff. Anyway, we used to have a lot of electronics in our dorm rooms at college back in the 1960s, but those were stereo systems.


How do you regard multicasting in networks? Any future use you might predict as bandwidth increases or do you think it's doomed to it's current role (discovery etc.)?

Predict that percentage of Internet traffic that is multicast will increase over time as news, entertainment, and information invade.


What do you think of darknets/cipherspaces such as tor, freenet, and i2p? From a technology standpont and an ethics standpoint.

The Internet needs less anonymity, because the first step toward savagery is the mask.


Hey bob my question is looking back on how much the internet has progressed over the years did you have a different purpose for it?

The Internet intelligentsia from the 1970s are outraged at the newbies who have dared to use the Internet for purposes completely unintended, like advertising, like YouTube. Tough. I cannot wait to see the next big new applications enabled by the Gigabit Internet. Connectivity is good.


Do you think someone else could have done it at the time?

Yes, but they didn't. I was lucky to be born to my parents, to accidentally get accepted to MIT, to sneak into Xerox Parc, and lucky to get the completely new problem of having a building full of personal computers, one on every desk, if you can imagine.


Thank you.

You are welcome.


What is your favourite color?

It is blue right?

The Internet's favorite color is transparent.


Look man, I'm on Reddit's IEEE 802.5 network, so like don't ASSume...

Anyway, why did you guys name it Ethernet and not something fancy like GodsBlood or MoreImportantThanAir?

You lie. Nobody has IEEE 802.5 networks anymore, even IBM has given up on Token Ring. GodsBlood and MoreImportantThank Air were already taken, so we went with a name that communicated omnipresent passive medium for the propagation of electromagnetic waves, starting with thick coax, but today wireless and on fiber.


I'm using the internet that Al Gore invented. Yours is rubbish.

Al Gore never claimed to have "invented" the Internet. No. What he claimed was to have "initiated its creation," which is completely different, especially in 1991, not 1969 when the Internet's (Arpanet's) packets first started flying. But I take your point.


Was token ring a better technology than Ethernet back in the day when both were in use?

No. Even though our beloved IEEE 802 standardized IBM Token Ring, having sold it myself, I can say it was never really open (had SNA dust all over it), and it was slow and expensive compared to IEEE 802.3 Ethernet. IBM never really got how to be an open standard during the LAN Wars.


Thank you, I hated token ring.

I'm with you on that.


I don't have anything to ask - I just want to say thank you!

You are welcome.


Hiya Bob.

I hate this. There are about a billion questions I'd like to ask about what how you saw the emergent technologies affect society, etc., where you see IP6 going, how you think Microsoft has positively or negatively changed the world, but all I can think is....

"Obviously a stooge for the W.B. Mason company. Wonder how much they paid him?"

Reddit has ruined me.

Am feeling myself sucked into Reddit.


What is your biggest pet peeve?

That nobody ever changes their mind.


What are your views on American government trying to censor and 'control' the way people use the internet?

Governments should leave the Internet alone.


How do you feel knowing that you had such a big impact on the whole world ? Is that fulfilling?

Mostly now I want to share the credit with the hundreds of people who have invented Ethernet over the last 40 years.


What was your protocol review / update process? How did you know which layers would change, especially when considering industrial production lines of different CPUs?

How can we use a similar approach to incorporate future network variables?

Thanks, by the way. I'm a network engineer and I really like looking at the bytes the packets consist of.

The layering of Internet protocols is its greatest invention. Layering has allowed me to live a rich full life at layers 1 and 2 while a bunch of other people got to play above me without permission. All that serendipity. Ethernet and TCP/IP were invented in 1973 in Palo Alto, where I am this second, and the World Wide Web was not invented until 1989 and the plumbing still worked. Whoa, dude!


It was recently brought to my attention that some people on the Internet believe that you can increase the speed across Ethernet by twisting more than one cable together, as seen here. Can you confirm this is entirely not possible?

Not had time to try it myself, but are you serious? Not possible would be my guess, not having tried it. Actually, I've know this for years, but have been keeping it a secret to make more money.


How much are you worth at present?

Do not believe everything you read on the Internet.


What are your feelings towards Bill Gates?

Great man. It would be very hard to find someone else better at being the richest man on Earth. Bill may not like me for going after Microsoft for its anti-competitive practices during the 1990s, but I meant no harm.


Have you ever worked with Sir Tim Berners-Lee?

Sir Tim holds the 3Com Founders Chair at MIT, where I am a Life Trustee, so I bump into Sir Tim now and then, like at SXSW when he came to Austin. He is like Gandhi with ADD.


Has Ethernet changed much since its creation?

Yes, quite a bit. For example, 2.94Mbps on thick coax, to 10Mbps on twisted pairs, to 100Mbps, to 1Gbps, to 10Gbos, to 40Gbps, to 100Gbps, and next to 400Gbpos and finally? 1Tbps. Also, gone wireless to WiFi and onto fiber for long-haul. Quite a bit. Many other inventors involved.


About 15 years ago I used to install Ethernet. Did you use an acronym to remember the color code?

Have long ago forgotten any color codes, sorry, except maybe ROY G BIV.


Well done sir. Thanks for your contribution to mankind.

You are welcome.


What conivinced you ethernet was possible to create when 10 yrs. before hand it was never even thought of?

The terminal on my desk at Xerox Parc was communicating at 300bps the day before we installed an Alto PC and CSMA/CD Ethernet running at 2.94Mbps, which is about 10,000 times faster. We went that fast because we could, and because our new laser printer could consume 20Mbps.


What do you think of Google?

Wish I had thought of that. As a professor of innovation, I like Google especially because of its "pivot" from fast search to auctioned targeted advertising. Google unseated Microsoft which unseated IBM. Who will unseat Google? Cannot wait to see how that plays out.


I'd love to hear your thoughts on all of the internet privacy things that have been going around for the past year or so now, such as CISPA and SOPA.

Private property is a great invention for innovation and economic growth. The Internet needs to deal with property properly.


[No question]

To join in the celebration of Ethernet's 40th Birthday -- gather innovation lessons, sing the unsung heroes, party -- go to http://www.facebook.com/Ethernet40thAnniversaryIEEESA


Wish more people understood just how versatile ethernet is. What are your thoughts on FCoE vs iSCSI when transmitting data for storage products?

Ethernet vs Ethernot usually ends up Ethernet. The Network Effect (as quantified by Metcalfe's Law) plus all the Ethernet infrastructure that has accumulated over 40 years. Remember, RS232C circa 1962 is still out there.


Hello Professor Metcalfe! I took a Computer Networks class this semester; one of the slides showed your sketch for the Ethernet :) The whole multiple-access idea and the protocols behind it is fascinating to me, and it's an honor to be able to speak to you. I want to ask you one of our final exam questions: What would happen if all of the Internet was simply an Ethernet with switches, and MAC addresses were used instead of IP addresses? I think I kind of flunked this question (said routing tables would be huge and mobility would be problematic), so I'm curious about your perspective. And one more question: do you think more efficient multiple access protocols can/will be invented in the future?

Have heard from those still in the packet plumbing industry that the trend is back from Internet routing toward Ethernet switching. Have no dog in that hunt. Ethernet was first designed to be LAN packet plumbing for the Internet. Have been predicting for years that TCP/IP/Ethernet would be replaced by some version of lambda switching, but that has not happened and I'm not sure what that means anyway. The trend was away from multi-access for a while, with the reemergence of wiring hubs, but then along came WiFi and LTE.


Why didn't you make it so that the packet preamble couldn't be reproduced in the packet payload? More broadly, what were the security considerations you had when designing the protocol?

Our early design of Ethernet assumed that security would be taken care of at higher levels of protocol. Ha!


Is this your first time on Reddit?

Yes, first time on Reddit. It's exhausting!


Ethernet has been around forever!

  1. When do you see it being replaced?

  2. What will it be replaced by?

  3. In hindsight, is there anything you would change about your input into 802.3?

Ethernet has been evolving and re-invented for 40 years, so I would not say that it's been around forever, nor that it will be replaced -- already has. The name has stuck. Much time is spent trying to decide what the word Ethernet means, to dispute whether I invented it or not. I think it has become a brand, a promise of openness and interoperability and high speed and low cost and preservation of the installed base and fierce competition among suppliers and rapid evolution of the IEEE standards following market engagement. Long live Ethernet!


What is your preferred color of Ethernet cable? A sweet royal blue? Canary yellow?

Yellow is the official Ethernet cable color, in my mind. I wonder if IEEE has a spec on that.


When the internet first became a thing, did you think porn would have such a huge part of it?

What?! Porn on the Internet? Ethernet filters out porn.


One of the cool memories I have from my time in the USAF was seeing an IMP being installed. If only I had known then...

Also, was there something used before Thicknet? Running that stuff around a building was challenging. I'm so glad twisted-pair took off!

edit IMP, not ICMP.

Thick coax was our initial choice because it could be passively tapped.


What are some future technologies you are looking forward to?

Detecting, deflecting, capturing, and mining asteroids. Actually, anything that Elon Musk is doing.


Where did you get the idea from? What inspired you to create the the Ethernet?

Arpanet (Internet 1.0) packet switching and Alohanet multi-access randomized retransmissions.


I read Metalcafé. Sorry Bob.

Huh?


How do I send data into the future?

Just leave data lying around, and it will get to the future automatically.


[No question]

Today we are celebrating Ethernet's 40th Birthday at the Computer History Museum in Mt. View, CA. We will be gathering innovation lessons from Ethernet history. We will be singing Ethernet's unsung heroes. And tonight we'll have a party. Happy Birthday Ethernet!


[No question]

Videos from yesterday's Ethernet innovation conference at Computer History Museum: http://www.netevents.org.uk/celebrating-40-years-of-ethernet-innovation-introduction Ahoy!


This interview was transcribed from an "ask me anything" question and answer session with Bob Metcalfe conducted on Reddit on 2013-05-21. The Reddit AMA can be found here.