Follow Slashdot stories on Twitter


Forgot your password?
Network The Internet Science

Caltech and UVic Set 339Gbps Internet Speed Record 79

MrSeb writes "Engineers at Caltech and the University of Victoria in Canada have smashed their own internet speed records, achieving a memory-to-memory transfer rate of 339 gigabits per second (53GB/s), 187Gbps (29GB/s) over a single duplex 100-gigabit connection, and a max disk-to-disk transfer speed of 96Gbps (15GB/s). At a sustained rate of 339Gbps, such a network could transfer four million gigabytes (4PB) of data per day — or around 200,000 Blu-ray movie rips. These speed records are all very impressive, but what's the point? Put simply, the scientific world deals with vasts amount of data — and that data needs to be moved around the world quickly. The most obvious example of this is CERN's Large Hadron Collider; in the past year, the high-speed academic networks connecting CERN to the outside world have transferred more than 100 petabytes of data. It is because of these networks that we can discover new particles, such as the Higgs boson. In essence, Caltech and the University of Victoria have taken it upon themselves to ride the bleeding edge of high-speed networks so that science can continue to prosper."
This discussion has been archived. No new comments can be posted.

Caltech and UVic Set 339Gbps Internet Speed Record

Comments Filter:
  • Progress (Score:5, Informative)

    by SternisheFan ( 2529412 ) on Wednesday November 28, 2012 @09:16PM (#42125353)
    The older neckbeard hugs his 300 baud modem, and softly sobs...
    • by Osgeld ( 1900440 )

      I remember being opposed to 28.8 modems cause the ANSI art from bbs's would scroll by too fast, it just wasnt the same without watching a menu draw on screen

      course now I use a machine routine to force my apple II to do 115,200 and thinking this seems a bit sluggish

    • No serious neckbeard used a 300 baud modem. The modem of choice was the Telebit Trailblazer.
    • Heh, I remember when I was younger connecting to 300 baud modems with my 2400 baud modem and thinking "man that's slow, it must be very old!" lol

  • My god (Score:1, Funny)

    by Anonymous Coward

    It's full of porn!

  • by nevermindme ( 912672 ) on Wednesday November 28, 2012 @09:27PM (#42125447)
    At this point it is just a matter of how much you are willing to spend with 10Gig Ethernet becoming the standard method for handoff from the Telco demark. Just installed a L3 endpoint with well over 50Gig/sec of capacity with each gig/sec CIR costing less than it takes my company to write the check and split it out to the customers invoices. Storage and Power/Cooling are the last expensive item in the datacenter.
    • Don't forget equipment management, billing, HR, insurance, and everything else it takes to keep your employees that actually install, maintain and upgrade that equipment around. Last I checked, people were still required and they are, by far, the most expensive part of our data center.

  • Also, if you want to bring next-gen gigabit fiber networking to homes in a major metro area your backhaul network needs to push the limits of fiber. Otherwise you run out of backbone capacity. Even with this speed, CDN endpoints are needed to reduce backbone bandwidth requirements for things like streaming video and TV.
    • I don't think these two institutions are concerned with streaming video and TV, or connecting to peoples homes at all. How many homes need to access 100PB of data from CERN?
      • Re:The point (Score:4, Interesting)

        by symbolset ( 646467 ) * on Thursday November 29, 2012 @02:46AM (#42127057) Journal

        I was taking a more general application of "what's the point?" The first connection to what would become the Internet was made between UCLA and SRI in Menlo Park, CA after all. That was a big deal for them, but a bigger deal for us. What the point of that was is rather subjective.

        100PB seems like a lot of data today - 3,000 times the 3TB storage available in a standard PC. But I am so old I wear an onion on my belt, as was the fashion in my day. 1/3000th of that 3TB is 1GB. I can remember when to have 1GB of storage in your PC was an undreamt of wealth of storage richness: a bottomless well that might never be filled. Hell, I can remember a day when 3TB of digital info storage was more storage than there was - everywhere on Earth. In fact in my early days there was talk of a terabyte being the sum of human knowledge (silly, I know). It's reasonable to expect that when that much more time has passed again, 100PB will not be a big deal either.

        So now we carry around a 1TB 2.5" USB drive in our shirt pocket like it's no big deal. And when guys like this do things like this we talk about what it means to them - and that's fine. But there is a larger story, like there was a larger story at UCLA - and that is "what does this mean to the rest of us?"

        Now 339Gbps isn't such a big deal. NEC wizards have already passed 101 Tbps - 300 times as much over a single fiber [], though not to this distance. That's enough bandwidth to pass your 100PB in 20 minutes, over a single strand of glass fiber.

        The LA Metro area is about 15 million people, or 3 million homes. To deliver 1Gbps to a reasonable half of 3 million homes and mesh it out for global distribution is going to require a lot of these links. The aggregate demand would probably be under 1% of peak potential of 3,000 Tbps or about 30Tbps. 100 times the bandwidth of this link. Using CDNs(*) - particularly for YouTube, CableTV, the usual porn suspects and BitTorrent you could diminish the need for wider bandwidth considerably but you still need a wide pipe to the outside world. And all the Internet servers in the world would need to be updated to support the crushing demand with higher performance, SSD storage and the like. And that's great for the economy, and it's just LA.

        These innovations are neat, but they're neater still when they come home to all of us.

        TL;DR: Get off my lawn.

        /(*) Define CDN: A CDN, or Content Delivery Network [] is a facility for moving high bandwidth, high demand or high transaction content closer to a nexus of consumers. An example would be Netflix, which delivers streaming video content to 21 million subscribers, comprising by some estimates a full third of Internet traffic. Netflix provides for free to Internet providers BackBlaze boxes [] that move Netflix content closer to the end user, reducing backbone usage. Similar boxes are provided by advertising networks and other content providers.

        • Re: (Score:3, Informative)

          by Anonymous Coward

          Now 339Gbps isn't such a big deal. NEC wizards have already passed 101 Tbps - 300 times as much over a single fiber [], though not to this distance. That's enough bandwidth to pass your 100PB in 20 minutes, over a single strand of glass fiber.

          Yes, but NEC sent raw bits (probably just pseudorandom data) over a single length of fiber, that's just the raw fiber throughput. This experiment ran over a live network, so: multiple lengths of fiber, optical amplifiers in between, add/drop multiplexers in between, then put Ethernet over it, IP routing between Ethernet segments, and _then_ terminate this on real machines transferring that data through memory to and from disk. The hardware they used is also commercially and currently available, while you'd

          • Standards are for weenies. If you own both ends of the glass pipe, how you put data down it is up to you. It's not standard. It's not Ethernet. But I really don't think Google gives a damn because they own both ends of the pipe and what they care about is moving bits for least cost at minimum latency.
            • Re: (Score:3, Interesting)

              by Anonymous Coward

              (same AC as above)

              That's true, but that ideal will break sooner than you'd like unless you're going to custom-design and build all your equipment yourself, which at that level is pretty tough, even for a company like Google. That equipment is not comparable to what you can get in your local computer shop, only very few companies worldwide can produce all devices you need from end to end.
              - After some 160km at most you're going to need amplification, which limits the wavelengths you can use (there are

    • by jbo5112 ( 154963 )

      Kansas City is already being wired with 1Gbps dedicated connections in people's homes. A fiber line runs from the home to a "fiber hut" (a room sized switch or router) and the fiber hut is placed on the Internet backbone, with no aggregation in between. I saw synchronous 900Mbit(ish) bandwidth tests results on a screen at Google Fiber Space. Even Verizon's and AT&T's 100Gbps network backbones are going to fill up pretty fast once this rolls out to more customers and Google starts installing for their

      • Google has the backbone fiber to do this, and the tech to make it work. Nobody else does. And they do, in fact, leverage non-standard tech to move their bits over their own fiber. They design their own switches, switch links and so on and they leverage emerging technologies like this. Just like they design their own servers, and go direct to China for motherboards and Intel for processors, they're not paying $1300 for a 10Gbps SFP+ LR GBIC.

        They bought this fiber for pennies on the dollar it cost to lay

  • So at home... (Score:3, Informative)

    by nurb432 ( 527695 ) on Wednesday November 28, 2012 @09:34PM (#42125509) Homepage Journal

    I could blow thru my bandwidth cap in just under a second.

    • by Anonymous Coward

      meanwhile comcast offers me 15mbit at 60 bucks a month

  • Inter-networking (Score:2, Informative)

    by fufufang ( 2603203 )

    For those who can't be bothered to read the whole article, the packets actually went over the Internet. It wasn't a simple case of direct optic fibre connection. It is impressive that the backbone can now achieve such bandwidth.

  • by Anonymous Coward

    What's the equivalent speed measured in Library of Congresses?

    • by mic0e ( 2740501 )
      You could transfer about one football field of library of congress area in the time it takes to watch one sitcom.
  • isn't that like the average home Internet speed in Iceland?
    • by Anonymous Coward

      meh, no, we only have 100 mbit/s standard, with expiriments being done on 1 gbit/s

  • Read the damn press release. It does say single link, but not single 100G link. If they were demonstrating in-line compression at those speeds, or actually sending already compressed data at 4x the line rate, that would be intresting news. Sloppy re-reporting.
    • It's good to know that 339 gigabits per second (53GB/s) is equal to 187Gbps and that's hows much iz gonna need next time I downloads 200000 Bluray films fore sundown. But how godamn many rods of horsepower am I'm gonna need to run this bastard. Them hamster gots to eat.

      Yeehaw, Iz ciphering real good now.

      Of course, even a dumb ole country boy knows that if the packets aren't routed, it's not fair to call it internet. I'll just chalk that up to sensationalist reporting by slashdot morons.

  • "... or around 200,000 Blu-ray movie rips."

    Does this score points with delinquent juveniles? It's a counterproductive addition that makes the whole posting appear specious with respect to academic and scientific relevance.
    • by sconeu ( 64226 )

      And the MPAA sued both universities for ONE BILLION DOLLARS, claiming the "potential to pirate 11.25 DVDs per second".

    • by Anonymous Coward

      Do you always stink that bad or is it only because you pooped your pants? Sorry to stoop to your level of trolling, but the summary at least gives both, the value in a reproducible measurement and as an analogy that relays how impressive the feat was. Why did you add your first sentence? At least I can blame you for mine.

  • Comcast and CenturyLink would still block you after using more than 250 gigs of bandwidth.

    In the future "unlimited" will mean "just a second".

    • Comcast and CenturyLink would still block you after using more than 250 gigs of bandwidth.

      In the future "unlimited" will mean "just a second".

      Australian so much loved here 100Mbit NBN has a 40GB quota :D Thats less than an hour full throttle.

  • by bragr ( 1612015 ) * on Wednesday November 28, 2012 @09:59PM (#42125665)
    >>Put simply, the scientific world deals with vasts amount of data This is so true. I work for a bioinformatics compute cluster and our users have no problem maxing out our infiniband infrastructure. Who needs load simulators when your users sometimes work with >1TB datasets?
  • "It is because of these networks that we can discover new particles"

    holy shit batman, I thought it was because of the "Large Hadron Collider", ok then +1 to you internet

  • That math is wonky (Score:2, Interesting)

    by Anonymous Coward

    How is 339 gigabits per second equal to 53 gigabytes per second?
    How is 187 gigabits per second equal to 29 gigabytes per second?
    How is 96 gigabits per second equal to 15 gigabytes per second?

    Glad to hear they're using 6.4 bits per byte. Next week they'll drop down to 2 bits per byte and push out another press release.

  • Sure, the LHC generates tons of data. But that will soon be dwarfed by some of the data sets that are coming from proteomics research, especially when you take into account how many proteomics labs around the world are (or soon will be) capable of generating 1TB /day.
    • To then be overtaken by the next generation radio telescopes, which will store TB/h as they test technologies for the SKA, which will store TB/second [].

      Seriously, no one has any idea how the fuck we're going to analyze all the data quite yet. A completely untrained n00b beats the pants off of any image classifying algorithm hands down, but how do you classify billions of objects that way?
      • foxconn. Duh

      • by jbo5112 ( 154963 )

        Facebook analyzes and stores roughly 500 TB a day (Apache web logs), just to know how their service is being used. I know it's quite an order of magnitude easier to analyze, but efficient cluster and distributed computing does wonders. Telescope data would fit the paradigm quite well, probably even playing nicely with the uber-simple mapreduce framework.

        Google figured out how to get untrained n00bs to classify images. They invented the Google Image Labeler game. IIRC, you would be paired up with someone

    • by Zondar ( 32904 )

      Earlier this year, Gizmodo reported that the LHC generates a petabyte of data every second. Actually, that's just one of the two primary experiments. [] []

      The have to preprocess the raw data through a filter that determines if the collision is interesting or not. Then they further parse that down and only store a tiny fraction of the raw data generated every day. If they were able

      • Re: (Score:2, Interesting)

        by Anonymous Coward

        That limit doesn't really have to do with offsite transfer speed, but just about local speed of getting it onto a disk a few meters away from the machine. Faster computers and storage technology is needed for that instead of faster long distance communication (unless a particular advance in the latter helps with dedicated lines over very short distances relatively speaking). Fast long distance operations improve the rate at which researchers can query and run tests on the large data set instead.

        Also, the

  • Surprises me they chose an island as the the northern terminus. I'm suspecting the undersea observation network has something to do with it.

  • by manu0601 ( 2221348 ) on Wednesday November 28, 2012 @11:28PM (#42126285)
    Do we know anything about OS and application setup? Achieving such speed is not obvious, and OS may kill copying data between user and kernel space.
    • by AmiMoJo ( 196126 ) *

      I'd imagine they are counting the raw network to RAM via DMA speed, not the speed at which the CPU can interpret the data. The disk transfer rate speed is a better indication of that, although presumably the disk DMAs the data into RAM and then the NIC DMAs it out again so the CPU is relatively unloaded.

      • DMA will not solve all your problems. If you use the POSIX API, for instance, you will call read(2) or recv(2) to get data in a buffer you provide. You do not know in advance the size of the data chunk, and if your buffer is to small, data ought to be waiting in kernel for you to retrieve it. Same problem if you call read(2) after the data arrived.

        In both cases you must have a copy between kernel and userland, regardless of how smart the implementation is with DMA

    • by Alarash ( 746254 )
      You are correct, it's not really viable to setup a whole bunch of PCs for this. If only to aggregate the results. You need to use dedicated load generators for that kind of testing. The aforementioned CERN uses this card ( []) for this. I know because we sold it to them.
  • Why is this news?

    In South Korea anyone can get 350Gbps fibre to their home for less than $50/month!

  • Doesn't it have to be both ways?

    (Like with land water and air soeed records...

  • 1 GB data cap and a cost of 399.99/month and you have to sign up for a trillion years ....
    who cares its does none of us any good with the restrictive long term copyright crap...
    it snot like anyone on the net is gonna be needing bandwidth unless your pirating right?

  • The faster we can move data is in some way connected to how fast life moves in every respect. There is a point when "cool" is truly a four letter word. Not that I am arguing against it, it's just that there is a similar arguments within the AI industry. In more and more ways our achievements are outstripping out ability to deal with the consequences.
  • I switched a guy over from copper DSL max 3mb/sec to our new fiber network, max TEN mb/s!

    UTMA []

  • My big question that the article seems a bit light on, is other than the "size" of the data, they make no mention of what kind of data was used.

    It's all very well and good if they used a randomized contents of a HD, then just replicated it. It is something else entirely if they used very specialized data for this particular test.

    i.e. that speed works very well but only if the data consists of all 1's.

    Also some pretty limited applications for this technology at present unless computers get faster, and storag

  • You should also mention the multi Petabyte paradigm of life scientists and the Next Generation Sequencing applications. This is just another frontier in data intensive computing paradigms.
  • This assumes a blu-ray movie is 20GB, which it is not. From Wikipedia []:

    "Conventional (pre-BD-XL) Blu-ray Discs contain 25 GB per layer, with dual layer discs (50 GB) being the industry standard for feature-length video discs."

    Hyperbole is good but should not be at the expense of the truth.

If you suspect a man, don't employ him.