Tomorrow's Internet: 1,000 times faster

Oct 21, 2010

Imagine if all the data traversing the world right now -- on long distance networks and between and within computers and other hardware -- could be sent through a single fiber the width of a human hair.

A new research center has been launched at the University of California, Santa Barbara (UCSB) to make that a reality. Researchers with the Terabit Optical Ethernet Center (TOEC) will develop the technology necessary for a new generation of Ethernet a thousand times faster, and much more energy efficient, than today's most advanced networks. They are aiming for 1 Terabit Ethernet over -- 1 trillion bits per second—by 2015, with the ultimate goal of enabling 100 Terabit Ethernet by 2020.

Partnering with TOEC as founding industry affiliates are Google Inc., Verizon, Intel, Agilent Technologies and Rockwell Collins Inc.

Internet traffic is booming, as businesses and institutions handle massive quantities of data and consumers stream video, share high-resolution photos and battle it out in online games. Millions of people will soon be consuming billions of bits per second in their living rooms, all at the same time.

"We're going to need much faster networking to handle the explosion in Internet traffic and support new large-scale applications like cloud computing," says Daniel Blumenthal, Professor of Electrical and Computer Engineering at UCSB and Director of TOEC, which is part of UCSB's Institute for Energy Efficiency (IEE).

"The work that will be conducted at TOEC will enable the future of the Internet," says Stuart Elby, Vice President of Network Architecture for Verizon.

Ethernet, the way computers talk to each other over a network, has become the de facto standard for data transmission both on a small scale and across global networks. "It's an accepted, flexible interface," says Internet pioneer David Farber, a professor at Carnegie Mellon University and former Chief Technologist for the Federal Communications Commission.

Ethernet is constantly evolving, but soon—in as little as five years, according to some estimates—it won't be able to keep up with the speed and bandwidth required for applications like video and cloud computing, and distributed data storage.

"Based on current traffic growth, it's clear that 1 Terabit per second trunks will be needed in the near future," Elby says.

Not only will Terabit Ethernet soon be needed to satisfy the demands created by the way we use networks now, but Farber says high-performance, high-speed Ethernet will open up opportunities we couldn't dream of today: "You build it, they will come."

"We are excited to be part of this new center and look forward to working with the other members to enable this future capability," says Mario Paniccia, Director of Intel's Photonics Technology Lab.

Current Ethernet technologies can't be pushed much past 100 Gigabits per second—the speed that's beginning to be implemented now—mainly because of the amount of power needed to run and cool the required systems, Blumenthal says. Large data centers can consume as much power as a small city. New generations of Ethernet need to be much more energy-efficient and cost-effective, or the power problem will limit Ethernet development, crippling the growth of key U.S. industries and technologies.

"Our goal," Blumenthal says, "is to make energy-saving technologies that will allow applications and the underlying networks to continue to scale as needed. You could think of it as greening future networks, and the systems that rely on those networks."

To achieve that, and drive Ethernet toward 100 Terabits per second, fundamental improvements in the underlying technologies will be required, he adds. "We're going to need dramatic breakthroughs across multiple disciplines, not only in the core Ethernet technologies but in Ethernet-based networking and in the engineering and measurement systems used to develop and test these new technologies," Blumenthal says.

Research at TOEC will build on UCSB's world-leading expertise in materials, advanced electronics, photonic integrated circuit technology, silicon photonics and high-speed integrated optical and electronic circuits, and in bridging these new technologies with real networking systems. Blumenthal says new low-cost, energy-efficient optical technologies that leverage the techniques now used in semiconductor manufacturing will be the foundation for the Ethernet of the future.

"Our strategy of using silicon photonics to create low-cost, integrated, Terabit-per-second devices," says Paniccia of Intel, "fits perfectly with TOEC's charter for energy-efficient high-speed Ethernet."

Doug Baney, Senior Research Manager at testing and measurement company Agilent Technologies, says "the characterization of the high-speed and complex nature of future Terabit Ethernet communications opens up whole new challenges in measurement instrumentation. That's a critical area that needs to be solved in tandem with component and system development to help pave the way to Terabit Ethernet deployment."

Explore further: Sensitive bomb detector to rove in search of danger

Provided by University of California -- Santa Barbara

4.5 /5 (22 votes)

Related Stories

Verizon completes industry-leading 100G Ethernet field trial

Jun 23, 2010

Continuing to lead the industry in 100G technology, Verizon has completed a field trial carrying 100 gigabit-per-second Ethernet traffic on a metropolitan Ethernet infrastructure. Using Alcatel-Lucent equipment deployed in ...

KVH launch

Aug 01, 2004

KVH Telecom (head office: Minato-ku, Tokyo, President and CEO: Rakesh Bhasin), a provider of integrated communications and IT management services, has announced that it will launch "Ether-MAN Plus," a next-generation Ethernet ...

Recommended for you

Sensitive bomb detector to rove in search of danger

13 hours ago

European researchers have developed and tested a light-weight device capable of detecting extremely minute quantities of explosives from up to 20 metres away, providing an invaluable law-enforcement tool ...

How soon could car seats enter the 3-D comfort zone?

May 23, 2013

New 3D textiles made of recyclable polyester fibres could contribute help cars be easier to recycle. But recycling technology has yet to progress in separating seat material from other car components.

NASA: Austin, calling Austin. 3-D pizzas to go

May 22, 2013

(Phys.org) —The idea of living with 3-D printed food is neither unthinkable nor new; designers and futurists have been looking to 3-D printing as food's next frontier. In 2012, there was news that the Thiel ...

User comments : 29

Adjust slider to filter visible comments by rank

Display comments: newest first

ThanderMAX
5 / 5 (1) Oct 21, 2010
Only increasing node-to-node ethernet speed will make routing/switching (specially in ISP side) even more difficult.

Even few 10Gbps or 100Gbps lines can saturate current (so called better-CISCO) routers. Unless those router changes it's architecture from current silicon backplane to light-switched architecture Internet will still be stucked in 2010's era, forever.
Noumenon
4.9 / 5 (52) Oct 21, 2010
Why only 1Terabit or 100Terabit and not 10000000000000000000000000000000000000000000000000000000000000000000000011110000100000000000000000000000000000000000011110001111000000000000011100000000000000000010000000000000000000000000010000000000001111000000000000000000000000000000000000000000000000000000000000000000000000000000000001111000011001100000000011100000000000000010000000000000000000000000000000000000000000000000000000000000000000000011110000000000000000000bits/s ????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????

Because of reality.
ziprar
1.1 / 5 (7) Oct 21, 2010
Its not the technology preventing high speed internet development but stupid ignorant goverments with nonsense reguliations and high tax.

Countries like south korea and japan (and some others) will definately benefit but certain others will lag even more.
trekgeek1
4.6 / 5 (11) Oct 21, 2010
Its not the technology preventing high speed internet development but stupid ignorant goverments with nonsense reguliations and high tax.



No, I'm pretty sure we have limits to our technology and need to figure some stuff out first. But good luck with your "government doesn't want us to have internet and water powered cars" paranoia.
Quantum_Conundrum
4.8 / 5 (4) Oct 21, 2010
One wonders what normal application could possibly require 100 terabit connection anyway.

Holographic video game maybe?

I don't know. The last 15 years in IT have been completely mind blowing, and the next 10 years are likely to be even greater.

It's hard to even fathom the technologies that are going to exist in 10 years.

I think back to the windows 95 era computers and I'm completely amazed that by 2020 we will have computers that are almost as far ahead of today's computers as today's computers are vs those.

What we are going to do with 100 terabits internet and 380 core processors I cannot fathom.
PinkElephant
4.6 / 5 (11) Oct 21, 2010
@ziprar,
stupid ignorant goverments with nonsense reguliations and high tax.

Countries like south korea and japan (and some others) will definately benefit
Wow, self-contradiction within 2 adjacent sentences. That's a feat I've only ever seen marjon accomplish. Bravo!

[For the non-cognoscenti, the high-speed networks currently in place in Korea and Japan, were built with taxpayer dollars by government mandate. Whereas countries like U.S. lag woefully behind even in high-density areas precisely because they emphasized private network development, while minimizing government involvement.]
DamienS
4.4 / 5 (7) Oct 21, 2010
For the non-cognoscenti, the high-speed networks currently in place in Korea and Japan, were built with taxpayer dollars by government mandate. Whereas countries like U.S. lag woefully behind even in high-density areas precisely because they emphasized private network development, while minimizing government involvement

Exactly right. Also it's easier to roll out a high speed network in densely populated, geographically small countries like Japan and Korea. Much harder in geographically larger countries like the US or Australia.

However, in the case of Australia, it is precisely taxpayer dollars (for the most part) that will be paying for their new nation-wide network rollout of fiber to the home (known as the national broadband network).
Graeme
not rated yet Oct 22, 2010
> However, in the case of Australia

So if a million Australians use their 100 meg Ethernet flat out, they will need a 100 terabit backhaul. Hopefully it can be done at a good price, and not in linear proportion to the speed.
Bob_B
5 / 5 (1) Oct 22, 2010
All I know is in the USA we'll be paying lots more for any improvement. Already many people are retreating from cable TV and using broadband entertainment...the price increases are on the way. If we won't use their cable and pay through the nose for it then we'll have to pay more (and more and more) for Internet service.
Vendicar_Decarian
3.3 / 5 (8) Oct 22, 2010
When Al Gore created what is now known as the internet by opening the existing military network to public access, he had no idea that 30 years of Republican Treason would conspire in a matter of just a few years to give the U.S. the worst network system of any first world nation, and many second world nations.

And in these dying days for the American state, we commend Al Gore for his vision and his lament that his country has proven itself unworthy of his good work.

Vendicar_Decarian
4.6 / 5 (9) Oct 22, 2010
"All I know is in the USA we'll be paying lots more for any improvement. " - Bob B

Correct.

The purpose of every American corporation is to extract as much wealth from the public as possible within the constraints of the laws that those corporations write for themselves.

finitesolutions
not rated yet Oct 22, 2010
It might be that the wireless technology will progress even further and render cables obsolete. Most likely quantum entangled particles will satisfy our communication needs without cables. But in the mean time you are lagging behind. Do not tell me that all the investment in modem technology has not payed off yet!
P.S. Everybody is entitled to a personal money making machine.
Eikka
5 / 5 (4) Oct 22, 2010
Soon every part of the world will be just 40 milliseconds apart. 80 ms if you want to get back again.

Latency is the biggest stumbling block with things like cloud services - not the available bandwidth. You can't deliver the same user experience if it takes you 200 ms to recieve information from the client and send the results back, because that means things like mouse cursor movement gets updated with a resolution of 5 frames per second, which is mighty choppy.

That's why multiplayer computer games run locally and only transmit the player positions, and even then they need to fake it by guessing where the other players will be and what they will do, because any action you take will take about 1/5 to 1/2 of a second to appear on other people's screens, and their reaction will be delayed by the same amount, so your game reality can be up to a second behind what's actually happening, and every player will run on a slightly different timeline.
Husky
5 / 5 (2) Oct 22, 2010
latency, good point, not only in games, but also important for real time distributed computing projects, Currently Boinc and the likes rely on distributing workunits that are processed offline on the clients and the results are then uploaded to some central server, but low latency would allow to really virtualize a seamless multicore processor, imagine the scientific possibilities if you could hook up several Crays and other supercomputers in different countries and access them as if they were one single Exaflop supercomputer
ashkot
not rated yet Oct 22, 2010
IS this available for countries like India where people still pays $232 for 512Kbps
DaveGee
5 / 5 (1) Oct 22, 2010
'this' isn't available at all... It's just the announcement of a multicompany work group that have tasked themselves with developing a TB network technology. Also, don't think for a moment that this is the only group working on faster networking tech... Heck google, intel and others are probably also working with a dozen or more like minded workgroups. After all there's no downside in being involved in as many groups as possible .... Early on .... It's not until there are two or three really close technologies when one has to stand up behind a specific tech.
knikiy
not rated yet Oct 23, 2010
I would be happy if today's internet was half as fast as my ISP says it is.
Quantum_Conundrum
not rated yet Oct 24, 2010
I would be happy if today's internet was half as fast as my ISP says it is.


Agreed. I'm supposedly on a 10megabyte (80 megabit?) per second connection, and to my knowledge, I can never download any faster than about 1 megabyte per second even on the lowest traffic sites, and usually it's much slower than that if you are downloading from a busy site. Torrents barely even make a difference in most cases.

The other thing is, 100TB would actually be 10 million times faster than the claimed specifications of my existing cable connection.
Buyck
5 / 5 (1) Oct 25, 2010
A 1000 times more faster internet... it was about time to lauch this initiative!
PinkElephant
5 / 5 (3) Oct 25, 2010
@Quantum_Conundrum,
I'm supposedly on a 10megabyte (80 megabit?) per second connection
I'd double-check that, if I were you. Typical cable offerings are 10 megabit, not 10 megabyte per second. That would amount to just a bit more than 1 MB per second, which is consistent with the max download rates you've observed...

That said, I'd also note that there's a difference between burst vs. sustained data rates. Typical consumer-grade connections are relatively cheap for a certain amount of bandwidth, because the ISP doesn't actually expect you to continuously pound your connection for all it's worth. So, they take 10 Gb/sec of bandwidth, divide among 10,000 customers, and give each customer a "10 Mb/sec" while assuming they won't all max out their activities at once... but when that happens (during peak hours), bandwidth per customer drops far below the advertised maximum.
Olivia
not rated yet Oct 25, 2010
Why only 1Terabit or 100Terabit and not 10000000000000000000000000000000000000000000000000000000000000000000000011110000100000000000000000000000000000000000011110001111000000000000011100000000000000000010000000000000000000000000010000000000001111000000000000000000000000000000000000000000000000000000000000000000000000000000000001111000011001100000000011100000000000000010000000000000000000000000000000000000000000000000000000000000000000000011110000000000000000000bits/s ????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????

I know you're funny. The amount of speed can enable teleportation, don't doubt its uses dudes.
CreepyD
5 / 5 (1) Oct 25, 2010
I think getting everyone onto broadband should be way ahead than getting a select few onto super fast broadband.
Modernmystic
1 / 5 (1) Oct 25, 2010
I think getting everyone onto broadband should be way ahead than getting a select few onto super fast broadband.


The law of the minimum. This is actually what "needs to be done" any system is constrained by how quickly the lowest aspect of said system is allowed to grow contrary to popular opinion. Systems like this are "grown" from the bottom up, not the top down.

I suspect many people will differ on the means to achieve this, but no the ends.
Skeptic_Heretic
not rated yet Oct 25, 2010
The law of the minimum. This is actually what "needs to be done" any system is constrained by how quickly the lowest aspect of said system is allowed to grow contrary to popular opinion.
Then why are you against the net neutrality act?

I'm just asking as a matter of clarification from stances you've held in the past.
Modernmystic
1 / 5 (1) Oct 25, 2010
All I'll say about net neutrality is that it's a very difficult and conflicted "problem" for me.

I see both sides of the argument, and I see merit in both sides.

However, to make a VERY long story short I do tend to lean towards net neutrality...at least this week.
Skeptic_Heretic
not rated yet Oct 25, 2010
All I'll say about net neutrality is that it's a very difficult and conflicted "problem" for me.

I see both sides of the argument, and I see merit in both sides.

However, to make a VERY long story short I do tend to lean towards net neutrality...at least this week.

Reasonable.

I tend towards a similar incllination. The bill as proposed isn't what I want, but there are some definite merits to some of the provisions therein.
GSwift7
1 / 5 (1) Oct 25, 2010
The fastest/cheapest internet services in the US appear to be government-provided services. For example, the city of chattenooga tennessee is soon to become the fastest in the US (supposedly), as far as home service.

http://www.dailyt...9616.htm

Cable companies have a hard time competing because they have to justify the expenditure to upgrade versus the time it takes for the infrastructure to become obsolete and therefore worthless on paper. A municipal system doesn't have to worry about obsolesence. It doesn't have to be paid for and produce a profit before it becomes devalued on an investment report.

It has more to do with competitive forces holding back hasty decisions than anything else. Businessmen are by nature cautious.
Gesus_Christ
not rated yet Oct 26, 2010
Holy Cow Dung!
I just jizzed in my pants...... I need more storage :S
VOR
not rated yet Oct 31, 2010
Its not the technology preventing high speed internet development but stupid ignorant goverments with nonsense reguliations and high tax.



No, I'm pretty sure we have limits to our technology and need to figure some stuff out first. But good luck with your "government doesn't want us to have internet and water powered cars" paranoia.


Its not paranoia. The US is far behind many countries in speed and availability. The reasons for that are a combination of capitalistic conflicts and governmental complications. So it may be that it will take technology to develop speed, but our culture will probably keep us behind in the deployment of that tech anyway.

More news stories

Google eyes emerging markets networks

Google has become deeply involved in a series of projects to build and operate wireless networks in emerging markets including sub-Saharan Africa and Southeast Asia, a report said Friday.

Facial-recognition technology proves its mettle

(Phys.org) —In a study that evaluated some of the latest in automatic facial recognition technology, researchers at Michigan State University were able to quickly identify one of the Boston Marathon bombing ...

Drones may violate international law

(Phys.org) —As President Obama gives a speech on national security—including defending U.S. use of drones to combat terrorism—Leila Sadat, JD, international law expert and professor of law at Washington University in ...

Galaxies fed by funnels of fuel

(Phys.org) —Computer simulations of galaxies growing over billions of years have revealed a likely scenario for how they feed: a cosmic version of swirly straws.