In the beginning, Ethernet was optional. When Bob Metcalfe and David Boggs cooked up their network protocol at Xerox PARC in the early 1970s, it was meant to connect the research haven’s now famous Alto machines — but only if researchers felt the need. “Each scientist would get a kind of Alto order form,” Metcalfe remembers, “ …
Bob Metcalfe did OK with Ethernet, but...
Bob Metcalfe did OK with 3Com, but everythig else he's touched he's managed to turn gold into lead. Everyone in Silcion Valley as well as Boston (where Metcalfe is based) knows that if he's involved with a start-up in any way, it's idea is probably dumb and certainly DOA. (Just look at Ember - pure crap!)
One trick pony.
Bob's a one trick pony who has managed to parlay that one trick into a persona who manages to find a means of separating fools from their money.
Funny thing is, that one trick wasn't his to begin with ... Another funny thing is that if he hadn't become involved, we'd have had WiFi a lot sooner. Maybe in in the early '80s, even ...
Side note: I've beaten Bob at tennis every time I've played him. Oh, and Bob, Linux is still here. My Win2K boxen are no longer relevant ...
Whatever. History is history, even when it's manipulated to provide "heroes" for idiots.
Ethernet dominated only because it was cheapest. I reckon Token Bus was the best technical bet soley because of its deterministic throughput such as you'd need for real time things like media streaming, network stability under heavy loads, etc. It was the VHS-Betamax thing, only in this case we never really saw the 'Betamax'.
Networked avionic systems on Boeings and Airbuses owe a lot more to Token Bus than Ethernet. They can't afford to fake real time network traffic by using gigabit - they have to do it properly.
a man at the cutting edge...
"Metcalfe went looking for a new Ethernet cable"
Wow must get one of these special ethernet cables. We just use generic RJ45 Cat5e's.
The National Grid, surely.
The problem with physics...
The problem with the physics it is no respector of a reputation gained in another area. The issues regarding communications would have been solved one way or another - if it wasn't Ethernet that won out, it would have been something else. It suceeded through a sort of Darwinian process, and just shows how all pervasive a technology can be if it is available cheaply and can be freely adopted by entrepreneur companies. Fundamentally, once the immense capital cost of R&D is done, the incremental costs of communications is very low.
The problem of power generation is a completely different issue. It's not just a matter, as it was with the communications revolution, of technologies emerging from other areas which presented the opportunity. The technological direction of semi-conductor production virtually guaranteed, that in a reasonably free market, it would get exploited in this way. It is almost impossible in a western market to suppress cheap, exploitable technology. It oozes out through the cracks of veted interests and fills the voids and suddenly you are in a different world.
There is absolutely no such obvious cheap technology direction with power generation. Well there was, but it's called fossil fuels, and the resources laid down over many 10s of millions of years are being used up at a colossal rate. Beyond that it gets difficult - the fundamental power of the universe is nuclear fusion. Almost all natural systems on the Earth are powered by it one way or another (tides and the inner heat of the Earth are exceptions). However, the technological exploitation of this in a cost-effective manner, whether directly, or indirectly through trapping the SUN's energy, weather systems, biofuels or the like is an extremely tough problem. There is no guarantee that if it is done, that it will be cheap.,
Stating there is a need for something doesn't mean it will happen if the method doesn't exist. We have a need for a cheap solution to health and social care. That is never going to happen (well maybe if the robots take over, but then they probably won't have much use for us).
At least he knows himself :)
"I was - I still am - naive about politics," Metcalfe says. "I didn't really understand the process. ....."
""Too many of the people working on this problem are luddites and greens and Marxists and politicians and lawyers and other people who don't understand the problem...."
"...we're going to take the next thirty years to break the back of the energy monopolies.."
I'm sure I'd enjoy working with this man, on purely technical projects; but he never even found his way to Politics 101, lol. Good luck to him I say, he needs it.
"Shortly thereafter, Metcalfe ran into an Intel engineer looking for new things he could build with the company's new PMOS manufacturing process. Naturally, Metcalfe suggested an Ethernet chip. "Next thing you know, I had DEC, Xerox, and Intel in the same room."
For a long time the network connector on the box was a 15-pin D Type known as a DIX connector (Digital Intel, Xerox). A transceiver module then plugged into this for whatever network medium you had, Twinax etc. Say what you like about the old stuff but those connectors never broke, unlike the plastic RJ45 connectors today.
Mine's the one with the coax stripper in the pocket.
Well he could have a point for his Enernet.
With current technology allowing Power over Ethernet and Ethernet over common power lines, I can easily see his idea gaining momentum pretty quick. Provided everyone (Companies controlling the wires) is willing to lend a hand in making the infrastructure compatible so that it becomes a mere hand gesture for the rest of us; plugging in and getting your power.
Combined with the new waves of low-power devices, his idea makes perfect sense.
Just thirty years?
Ethernet is a very good example, in the wider sense, of why "open" projects win. I don't know exactly how open Ethernet is, but certainly it's sufficiently open that just about any company that wants to build ethernet-connected networking hardware, gets to do so. They compete and co-operate at the same time. The best ideas become standards, the others wither on the vine.
The First Ethernet I used was a yellow co-ax cable about 15mm diameter, into which you drilled a little hole to access the 10Mbps signal conductor in the centre, and onto which you clamped a transceiver measuring about ten by four by two inches. That was replaced by skinny co-ax like a TV aerial. That in turn by cat-3 cable with RJ45 at 10Mbps, then cat-5 100Mbps, then 1000Mbps. Then fibre, of various sorts, if you wanted faster or further. The original "ether" has been all but completely replaced by point-to-point communications and active switches. But it's not just the name that survives. The old stuff continues to work and interconnect for as long as anyone wants to keep it powered up. We still have an ancient piece of very-expensive-to-replace lab equipment connected via skinny co-ax ethernet and a coax-to-10BaseT hub, to a mostly Gigabit switched network. It works just fine.
why everyone seems to have a "dark little heart" ? ;-)
Paris, coz she also had a little dark corner which may not be so little anymore.
Yeah, what a failure.
All he did was to create Ethernet. Tosser.
Does not make me nostalgic
I've played find the loose Ethernet co-ax connector at 2am in the morning. Oringinal concept (inductive linke to backbone by transformer) robust and used by avionics buses (not sure if it pre- or post-dates 1553b). Cheap version with T connectors rather less so. But very pervassive.
Heard Raytheon UK was still using this < 5 yrs ago however. Very easy to cut blocks of PCs off a network, stitch it back together afterward. Handy for those high security defense software projects I imagine they carry out.
But it is amazing. From 10mbs to 1gbs on copper wire. This never would have been believed when originally conceived.
Nostalgia is the universe's way of telling you large parts of your long term memory are decaying.
BTW Did he actually get his Phd? Or does he join bad boy Billy as the worlds second most sussecful programme drop out?
Ethernet won because it worked better in real life than the alternatives, despite looking worse on paper.
Token Ring and similar systems work like the meetings in "Lord of the Flies": everyone sits in a circle, and no-one is permitted to speak unless they have a token (the conch) in their hands. The need to pass the conch around the circle from one party to the next, always in the same direction and including even people who have nothing to say, creates unnecessary delays.
Ethernet works like a dinner party conversation: you just speak and if you can't understand what you're saying because someone else is speaking at the same time, both of you shut up for a random period of time before continuing from where you left off. Sound like a recipe for disaster, with delays while collisions are resolved surely taking longer than passing around a conch?
In practice, as long as each person waits a *different* amount of time before reopening their mouth, it works very well, and the delays caused by the inevitable collisions are still less than the delays caused by passing the conch. With Cat5 cables and modern switches (as opposed to old-fashioned hubs or even older-fashioned co-ax cables), you can even have several separate conversations going on at the same table at the same time: once you've established who you are talking to, you can move up close to them, whisper straight in their ear and not disrupt what anyone else is saying. The only time you can get a collision is when you have to shout for someone else's attention.
Lesson: In theory, theory and practice should be the same. In practice, they often aren't.
"IBM had no intention of creating a standard," Metcalfe says. "But in their dark little heart they didn't really believe in it".
Token Ring had more "logic" but was too heavy physically, and yes I believe there was plenty of "small dark hearts" at IBM at the time. (I am a hell of on optimist, I suppose).
"Technological innovation is the source of all progress" yes, sure, starting with the stone axe, for instance.
But certain conditions are needed to enable "Technological innovation", I think we should remember that too.
We Already Know the Way
Abundant and clean energy, but not necessarily cheap, can be provided by nuclear power, so that we can continue to survive at a high technical level as would be required to be able to afford to pay for the research needed for either solar power satellites or controlled thermonuclear fusion (or, preferably, both).
But this wouldn't "break the back of the energy monopolies" the way warm and fuzzy technologies like windmills and geothermal and tidal power would.
We're just lucky that Thorium 232 can be turned into Uranium 233, and Uranium 238 can be turned into Plutonium 239, as easily by neutrons from Uranium 233 and Plutonium 239 fission as by neutrons from Uranium 235 fission, so that the delay in moving to utilize common nuclear fuel resources instead of only scarce ones will not lead to disastrous waste.
So unless you keep doing really cool stuff you're a twat?
These comments are great, so Bob;
. Turns gold to lead.
. produces "pure crap"
. is a con artist
. is guilty of plagiarism
. held up WiFi
. can't play tennis
. doesn't believe in linux
. produced inferior standards
. does not understand politics
. is irrelevant as it would have been someone else if it wasn't him
Just listen to yourselves, it doesn't matter if what you say is true or not, you sound like petulant children, knocking the man doesn't demean the pervasiveness of ethernet, the best one of you underperformers can manage is "he did OK with 3Com" OK? OK? WTF? that's like saying "McDonalds did OK with that burger selling thing".
In my considered opinion, you are wankers (not in a good way, you have made a wank-fan of pictures of yourself and are furiously spanking as your mum and her friend from church enters your bedroom, to be met by a milk truck crash at the billy mills roundabout).
Have a word with yourselves.
When I read the byline:
When ten people stood up to ask 'What happened?' Metcalfe realized 'Ethernet was not an option'
I thought it was a quip on collisions.
"Wow must get one of these special ethernet cables. We just use generic RJ45 Cat5e's."
In many (I'm even tempted to say 'most') large electronic retailers in the US Cat5e and Cat6 network cables are labeled as "Ethernet Cable" and only occasionally do the cable's actual type or spec appear anywhere on the package. In keeping with the general hiring policy of these sorts of stores the dim-witted high-school drop out sales people will generally respond with either blank [dummy-mode] stares or confused face scrunching and head scratching if you ask for a "Cat5e cable" or anything that doesn't exactly match the label their company puts on the inordinately marked up and rebranded Belkin patch cables.
Noone done this yet?
One Server to rule them all
One Ping to find them
One LAN to bring them all
And in the darkness BIND them
Tolkien Ring Network.
Mine is the one with the orc.
@Mike re Haters
Got to be FOTW. Awesome post
The world needs more people like Metclafe
...knock him all you want, but decades later Ethernet is used everywhere and it works. I work in the renewable energy sector and it boils my piss having to deal with vested interests and obfuscation. We need people like him to cut through it.
As ever, PARC set the historic precedent ...
... for the way decisions are still made in the computer industry to this day.
>"When ten people stood up to ask "What happened?," Metcalfe realized his fledgling network protocol might be a keeper. "From then on," he says, "Ethernet was not an option.""
"Hey! It's flakey as hell owing to a bad design and barely works! Ship it!"
In any sane field of engineering, discovering that something can be so trivially broken would be a reason to roll it *back*, not roll it out. Oh well.
@A J Stiles
Ah, love that description of networks - takes me back to my school days. You might want to add the "if two people next to each other fall out, nobody gets to talk" for the token ring analogy!
But as the BOFH once said, "show me an ethernet collision and I'll show you a network that could do with one user fewer"
A helper in their midst?
Back in '81-'82, before I became an engineer, I was a corporate sales rep. at ComputerLand in Los Altos, CA. In that capacity, I got to know Robert Metcalf and Howard Charney at 3Com when I sold them their first 100 IBM PC's. I later met up with Robert in Boston at some IEEE meetings and dinners. I will always admire and respect him and his straight-forward manner. He is truly someone who can tell when "the emperor has no clothes"!
Metcalf is passing tokens these days
Allowing for a wikipedian approach to fact checking, I'm pretty sure that not much of the original 'Ethernet' technology exists in what we gaily use on a day to day basis. If memory serves CSMA-CD was under severe strain with the 100Mb standards and was done away with for 1Gb. They replaced it with a 'token passing algorithm'.
So Token Ring won by changing its name to Ethernet.
Don't let that James Watt bloke try to fix your steam turbine either.
I enjoyed the article - but I had the same double-take as Stu Reeves at the mention of an "Ethernet" cable.
@ Mike again
So what you're saying Viz-a viz, is stratight out of the original Roger's Profinisaurus. I think a translation for the uninitiated is required.
Suffice to say, these people tend to get interrupted by their mums!!
...could be done with compressed air and last time I checked tou could keep it in a container indefinatly. http://www.youtube.com/watch?v=Dq8aZVLpf-c to see how air can run anything from generators to cars. Oh and if you happen to have a desert or see near you, you can generate compressed air. Eathernet enoth for you
Token ring vs Ethernet
Back when I worked for a trading bank, we could remote boot and access all our data over a 16Mb Token Ring network, whereas the side of the bank using 100Mb ethernet had serious problems just accessing data and people often had to work on files locally.
Just one little remark to the energy network idea: He is talking about active switched and buffered energy transmission. This is sometimes called a smart grid. It works by accounting for every amount of energy put into and taken of of the system and directs the required amount from the available sources to the destinations that need it, while buffering any exessive amount for later use. This would work like a big distributed battery, which can be filled when we have energy and drained when there is a need. This would allow distributed renewable sources to be used without unbalancing the network and also allows energy reuse, like regenerative braking in electric wehicles. I think this is a good idea.
"In any sane field of engineering, discovering that something can be so trivially broken would be a reason to roll it *back*, not roll it out. Oh well."
Quite right , in fact we shouldn't have shared electricity lines either , thats obviously a broken model. Every appliance should have a wire going direct to the power station!
10base5/2 ethernet actually was very efficient as far as cabling goes. You didn't need a whole roomload of racks and racks of switches that you do today with 10baseT. A entire office could be fed off one port and if it was cabled up properly - ie the co-ax not trailing across the floor and under desks - then network breaks were never an issue. Why 10baseT because popular I'll never understand.
Cheers Snake Plissken
Hadn't heard that one.
Token Ring vs. Ethernet
Ahh, but with your 16Mb/s token ring installation using mechanical MAU's, woe betide you if someone in the building attached a 4Mb/s device. Instant mahem, with no easy way to isolate the problem, especially if you did not know the order of the desks on the 'ring (the dreaded 'beaconing' error message.)
This took out a major UK IBM support centre for a whole day (we had to rely on 3270 over co-ax using MYTE on the few desks that had PS/2's on them). We then were tasked with splitting the ring in two, with a bridge tying them together to make a single network, so that if the same thing happened again, only half of the desks would loose their network connectivity (and we had to duplicate routers, 3270 gateways, printers, etc.)
Mind you, trying to work out what happened when someone unplugged the terminator of a 10base5 ('thick wire') or 10base2 ('thin wire') Ethernet had it's own problems as well!
And does anybody else remember the 'jabber' light on early 10baseT twisted pair hubs (not switches). Often caused when someone plugged an RS232 terminal into a structured cabling port that was connected to a hub.
I always felt uneasy whenever I had to tap into a 10base5 thick wire cable. It just felt wrong to take what was effectively a drill (the 'tap tool') to the cable that you had spent so much money and time having installed.
Thank god we now have intelligent switches, with automatic isolation of noisy/broken cables, and proper switch mode to eliminate the need for CSMA/CD.
RE: One trick pony.
Yes, Jake, but that one pony was one heck of a trick pony! I'd be very happy if I could claim something like Ethernet as my own. I can't imagine the fun and games of trying to get DEC, Intel and Xerox to agree a standard, let alone GM and IBM, so kudos to Metcalfe for finding the common ground. I wouldn't like to try and repeat the process today! I still remember the mess of early 100Base-T and fibre networking, where everyone jumped the gun and we had non-compatible technologies like hp's 100vg and FDDI versions. At least the base work of "open" Ethernet has allowed a solid build into 100Base-T, Gigabit and now 10GbE. I was a bit worried about the latter, I thought CISCO might get huffy and try and protect their marketshare by derailing 10GbE as a standard.
10 base T, and later 100 base T, became popular because it is faster than co-ax.
With co-ax, or unswitched 10 base T hubs, every packet gets sent to every host -- whether it wants it or not. The bandwidth is effectively shared between all the machines on the network; only one packet can be sent at a time.
10 / 100 base T switches can inspect the headers of incoming packets, and automatically route them only to the appropriate port. This isolation reduces collisions. An 8-port switch can potentially route up to 4 individual packets at any one time, and a 48-porter can route up to 24 packets at a time! Obviously, the first packet of a conversation has to be broadcast to all ports; but as soon as someone responds, the switch "learns" which port the destination machine is plugged into. From then on, packets for that machine are only ever sent to that port.
@ Mike Re: Haters
I vote FOTW too!
Paris, because we'd all love to have her as an old flame...
awesome post, love the comment about the billy mill roundabout, obviously another viz fan! I actually live only a few miles from the actual roundabout, and smirk everytime I go round it!!
How did Metcalfe and crew settle on such a crappy connector? Entire civilizations could have been built by redirecting countless millions of manhours effort devoted to untwisting, rearranging then stuffing wire pairs into connector bodies, finally attempting to hold the whole mess together while squinting and crimping. Couldn't there have been second, slightly less squishy connector option, where one --knows-- one has an actual connection and where the physical locking of the connector does not depend on the elasticity of a little plastic tab?
"So unless you keep doing really cool stuff you're a twat?"
No, Mike. The point is that if you're a twat, you're a twat. For some reason, humans of this era seem to enjoy turning twats into heros. They aren't heros. They are twats.
If I were an AOLer, I'd select the Paris icon, for what should be obvious reasons.
@Steven Jones As they say in lolcat world, Dinosaur, u iz one. Can't wait until you are our fossil fuel.
@Mike: FOTW for sure. I don't know anything about the billy mills roundabout, but, I'm going to use "Have a word with yourselves."
Wish they'd used a different name
Okay, so maybe light, and other electromagnetic waves, actually travel through a lack of a medium, but that not-medium has been known as "the ether" since the days of spark-gap.
Ethernet should have been the name of an over the "air" (radio wave) way of networking.
Metcalfe and Boggs could have called their creation "PacketNet" or something like that.
Besides, everyone knows that Ethernet packets are really delivered by the Ether Bunny. :-)
Mine's the one made out of recycled frozen yellow garden hose cladding and broken cellophane grass bits.
Jeez mate, what did Bob do to you - drown your favourite puppy or something?
From where I sit, he invented Ethernet, ripped it from the dead hand of Xerox and turned it into a very successful company (which seems to have been a decent place to work, based on the folks I know that worked there). Either of those would constitute a major achievement for most of us, but then I guess we're not in your class.
OK, he took the money and ran, but rather than buy a small Caribbean island he still involves himself in the industry. What's so wrong with that?
"Jeez mate, what did Bob do to you"
Nothing. I've known him since (about) 1976, which no doubt colors my perception. I haven't talked to him in probably two decades ... but I did read his columns in InfoWorld throughout the '90s (they may be available online, but I can't be arsed to look). Why read 'em? Water-cooler material. And occasionally a preemptive fact-gathering mission when he was dead wrong about whatever it was he was blathering about.
"What's so wrong with that?"
Again, nothing. More power to him ... But don't equate him having one "ah-HAH!" lightbulb moment with him being a genius, or even hero material. He's not. John Draper (who I also knew, back in the day) is far more worthy of "hero" status, nutter that he is ... although he'd most likely turn it down.
Metcalfe and Contributions
OK. First, a disclaimer: I know Bob Metcalfe. He was the Chairman of the board of a company I founded. He's still a member of the board.
Suggesting that his contribution was inferior to other approaches, or "didn't work" or whatever is patently absurd, but it also misses a major point: Ideas are cheap -- lots of people have interesting ideas, from the guy who wanders up and down the street talking about free energy from his inter-stellar visitor friends to Nobel prize winners. Ideas are cheap. Turning them into reality takes more than most people have: ambition, drive, resilience, and often charisma.
Ethernet at the start was not a slam dunk. (I have a copy of the note from Robert Bachrach at PARC completely trashing the idea of Ethernet and CSMA.) Bringing three different vendors together was not a slam dunk. Building a new company is never a slam dunk.
Folks who build stuff take risks. Some of the risks may seem DOA in hindsight. Success is never a slam dunk.
However, posting anonymously is always a slam dunk: nothing risked. (Unless your mommy happens to open the door.)
"How did Metcalfe and crew settle on such a crappy connector?"
Too young to have seen the connectors in use before 8P8C? Have a gander:
Old times: http://en.wikipedia.org/wiki/10BASE2
Faster but trickier: http://en.wikipedia.org/wiki/TIA/EIA-568-B
You wrote "Slam Dunk". Several times. As if it were an important phrase.
IMO, you can be safely ignored. I feel sorry for your mommy.
It didn't look like a very provocative article to me. Shows what I know.
Mine's the one etc ...
@Destroy All Monsters
Too young to have seen the connectors in use before 10Base2? Have a gander:
"at a high technical level as would be required to be able to afford to pay for the research needed for either solar power satellites "
What research are you thinking of? Mfg techniques, including how to use microwave oven grade magnetrons (very cheap but not considered precision frequency sources) were all worked out in the 70s, 80s and 90s.
OTOH if you mean lowering launch costs by 2 orders of magnitude that requires skilled engineers dedicated to the idea with a belief its possible and an adequate but not extravagant budget to try. That's engineering, not science. All know strategies to build an SPS fail on economics. Roughly speaking launch requirements give systems with very long payback. Reasonable payback needs big launch. The Iridum sat network cost c$5Bn to launch. Its total mass (66 sata @c1500lb ea is < 17000Lbs)is nowhere near a relatively small SPS, its too low and will still need replacing.