Bob Metcalfe thinks 1 Tbit/s Ethernet is inevitable, and he also believes the industry will have to tear down some standards to get there.
The "Father of Ethernet," as he's often called, will be delivering one of the keynotes at next week's OFC/NFOEC show in San Diego, sketching what he believes will be the path to get to 1 Tbit/s speeds.
And in an exclusive LRTV interview, Metcalfe points out that the path might leave behind the equipment and even the fiber that's been in the network for years.
"There comes a time when standards have to be overthrown," he says. "We're going down into sort of a dead end. I think that dead end is deep enough that we'll get to 100 Gbit/s. The evidence is, to get to terabit Ethernet, we'll have to break out of that dead end."
Metcalfe, of course, has no problem with that kind of thinking.
"There's now room to break loose of the stranglehold of standards and now move into some really fun new technologies," he says.
But would anyone even need terabit Ethernet? It's a question Metcalfe dismisses with his usual candor:
"We build it; they will come, I'm sure. It's happened every time for 35 years."
Interested in learning more on this topic? Then come to Ethernet Expo: Europe 2008, the premier event covering the hot topic of carrier Ethernet network technologies and services in Europe, the Middle East, and Africa (EMEA). To be staged in London, April 14-16, admission is free for attendees meeting our prequalification criteria. For more information, or to register, click here.
This guy Metcalfe sound like another "standards weenie" who gets hired by some company based on reputation and ends up screwing off and leaving the real work to the schleps back in the office while he get paid 5x what they do and racks up a huge expense account.
you know what... I'm reminded of when Larry Ellison brought out the "network computer" idea in the '90s. It was basically a dumb terminal -- a cheap machine that would tap applications that reside in the network.
Part of his motivation was to fight the bloat of Microsoft Office, and his presentations would include long screeds about how so many useless functions were piling up in Word.
Not that I'm a huge Ellison fan, but ... he really had a point there.
I'd wager he's not as enraged about flashy Web ads, though, since you can't blame it on Microsoft.
It was just not the advent of the PC. the devolution of the mainframe and terminal connectivity started with the mini from DEC and others. DECNet was really Ethernet in disguise. The focus was distributed CPUs connected to central processing vs. terminals connected to central processing. Ethernet won because it became an open standard that PC manufactures could adapt and exploit at cost with margin.
Token Ring lost because of the centralized nature of the architecture.
Bob is only pushing buttons in the market as he has done so well over the years. Terabit Ethernet will come; but the physics are still years away.
So, going back to the beginning of time. Ethernet came out and fell on its face. I remember working on Ungermann-Bass gear connecting some Data General Minis together with Ethernet.
It was much simpler to use serial interfaces for our need and we abandoned Ethernet.
What really got things going was the push of PCs into the workplace. Workgroups were set up independent of IT organizations. These workgroups used Ethernet instead of Token Ring, as they had no access to the Mainframe. IT ignored these, and they spread like wildfire. Then IT realized that they could grow their budgets by taking over these networks, and the Token Ring/Ethernet wars occurred.
Mainframes became less of the focus and Ethernet won that battle after several years.
Then came 10BT, 100BT, and GigE. We have not really seen 100GigE yet and the issues of analog nature of high speed technology have yet to be addressed.
So, if they build it they will come did not work for Ethernet. Something else changed and then Ethernet became viable.
Yeah, I don't like all these ads either. Neither did many in the public during the early days of RF broadcasting.
http://earlyradiohistory.us/sec020.htm
In March and April, 1922, Commerce Secretary Herbert Hoover sponsored a national Conference on Radio Telephony, which in part addressed the question of radio advertising. During the meeting Hoover warned that "It is inconceivable that we should allow so great a possibility for service to be drowned in advertising chatter", and the conference recommendations for advertising standards would have restricted it to near non-existence. The final report called for "toll broadcasting" to be the least important of four categories of stations, with limited transmitting ranges, and their development kept under "close observation". Moreover, commercial messages were to be "indirect" only, and "limited to a statement of the call letters of the station and of the name of the firm responsible for the matter broadcasted". The conference report, however, was never adopted as official policy, and a year later, the report of the second national conference did not include any restrictions -- or even references -- to toll broadcasting. However, the industry continued to cast a wary eye on developments, and at the third conference, in 1924, Hoover famously warned that "if the speech by the President is to be used as the meat in the sandwich of two patent medicine advertisements there will be no radio left". However, he added that "The listeners will decide in any event. Nor do I believe there is any practical method of payment from the listeners."
Ahhh... the old 555 timer... you're dating yourself! I admit that my original comment may be a tad cynical... but I'd like to see some conservation in terms of bandwidth (and memory)... I really don't like full motion video ads on my text web pages. Cheers to all!
Seems to me that the optical components needed for serial Terabit is just the tip of the iceberg. (I assume serial otherwise it's warmed-over DWDM)
What will the Ftau of transistors need to be to serve such critical functions as modulator drivers at terabit speeds? Anybody making such parts? How are they packaged? Will they cost as much as the optics given that volume will likely be low?
Assuming one can get there, in seems that only the Infinera's of the world are in a position to address the requirements because heavy integration of both optical and electronic functionality is a necessity. As for a new fiber infrastructure; somebody must be smoking some good stuff!! Terabit over copper anyone?
The path to Terabit is paved with parasitics and too little green! Place your bets folks!
The blogs and comments are the opinions only of the writers and do not reflect the views of Light Reading. They are no substitute for your own research and should not be relied upon for trading or any other purpose.
To save this item to your list of favorite Light Reading content so you can find it later in your Profile page, click the "Save It" button next to the item.
If you found this interesting or useful, please use the links to the services below to share it with other readers. You will need a free account with each service to share an item via that service.
Digg
Del.icio.us
Reddit
