Slashdot Mirror

← Back to Stories (view on slashdot.org)

University of Illinois Transmits Record 57Gbps Through Fiber Optic Lines (digitaltrends.com)

Posted by BeauHD on from the that's-a-lot-of-data dept.

An anonymous reader quotes a report from Digital Trends: Engineers at the University of Illinois have set a new record for fiber-optic data transmission, breaking previous theories that fiber optics have a limit in how much data they can carry. The engineers transmitted 57Gbps of error-free data at room temperature. The group, led by Professor Milton Feng, improved on its previous work in 2014, when it achieved 40Gbps. The keywords here are "error free," which is what makes this research unique from others that claim faster speeds. Fang said, "There is a lot of data out there, but if your data transmission is not fast enough, you cannot use data that's been collected; you cannot use upcoming technologies that use large data streams, like virtual reality. The direction toward fiber-optic communication is going to increase because there's a higher speed data rate, especially over distance." Engadget writes in an update to a similar report: "Reader Tanj notes that this is specifically a record for VCSEL (vertical cavity surface-emitting laser) fiber, not fiber as a whole."

57 comments

  1. What was the content by Anonymous Coward · · Score: 0

    Paris Hilton and uncredited male friends?

    1. Re: What was the content by BarbaraHudson · · Score: 1

      We know it wasn't virtual reality streams, which they say will need this. Don't believe it - the future is all reality shows and reruns.

      --
      "Transparent" is a shit show that trades on every stereotype going. A man in drag is NOT a transsexual.
    2. Re: What was the content by ArmoredDragon · · Score: 1

      Even if it was for VR, I suspect it would be very easy to get away with literally streaming your game from scratch with only a 50mbit downstream link. They could always adopt the approach Blizzard uses to allow WoW to be played long before you finish downloading all of the game assets, and with a 50mbit link, it would be totally seamless.

  2. Standard units please by sjbe · · Score: 4, Funny

    Please convert to standard units Library Of Congress's transmitted per micro-fortnight

    1. Re:Standard units please by Anonymous Coward · · Score: 0

      That's like trying to measure a bacteria with a meter stick.

    2. Re:Standard units please by Iamthecheese · · Score: 1

      Going by the book definition of 10 TB. (conveniently bypassing the question of digital archives) According to Wolfram Alpha a microfortnight is 1.21 seconds. 57 GB/s is 68.97 GB/mFn. Divide by 10,000 and you have:

      .006897 LoC/GB/mFn

      --
      If video games influenced behavior the Pac Man generation would be eating pills and running away from their problems.
    3. Re:Standard units please by Anonymous Coward · · Score: 0

      As long as your meter stick has nanometer markings on it.

    4. Re:Standard units please by Sir+Holo · · Score: 1

      As long as your meter stick has nanometer markings on it.

      If you know what material your meter stick is made from, you can use it to measure distances on the nanometer scale.

      But only over short distances. Tiny temperature fluctuations (just touching it) will change the length of the meter stick due to thermal expansion.

    5. Re:Standard units please by Anonymous Coward · · Score: 0

      mFn is a millifortnight, not a microfortnight. You want Fn.

    6. Re:Standard units please by Anonymous Coward · · Score: 0

      You want uFn, since Slashdot can't handle a mu.

  3. So much slower than 100GBase-RZ...? by ffkom · · Score: 4, Informative

    Excuse me, but even after reading the linked article it eludes me how this is an advancement over existing technology like 100GBase-ZR EtherNet lines (operating at ~ 120 Gbaud per fiber)?

    1. Re: So much slower than 100GBase-RZ...? by Anonymous Coward · · Score: 1

      Yeah, I thought we were way past the Tbps mark
      http://www.extremetech.com/extreme/192929-255tbps-worlds-fastest-network-could-carry-all-the-internet-traffic-single-fiber

    2. Re: So much slower than 100GBase-RZ...? by Anonymous Coward · · Score: 0

      The Tbps links are achieved with dozens of "colors" or wavelengths multiplexed together. On top of that, the guys from the story you linked used multiple fiber cores for spatial multiplexing (which in my book is cheating if you want to call it "single fiber"). This story is about a transmission with just one laser, so no multiplexing at all.

    3. Re:So much slower than 100GBase-RZ...? by Anonymous Coward · · Score: 0

      it uses VCSEL (vertical cavity surface-emitting laser), which has some advantages over other technologies.
      sales pitch here: http://myvcsel.com/key-advanta...

    4. Re: So much slower than 100GBase-RZ...? by ffkom · · Score: 1

      But still, the above linked 100GBase-ZR specs make use of only one "color" and a single fibre - for a higher data rate.

    5. Re:So much slower than 100GBase-RZ...? by Anonymous Coward · · Score: 0

      100GBase-ZR is 4 emitters multiplexed on the same line.

      This is done with a single emitter that is cheaper to produce in large quantities.

    6. Re: So much slower than 100GBase-RZ...? by Anonymous Coward · · Score: 0

      That uses "dual polarization quadrature phase shift keying". I haven't read the article, but I think the difference is that they are not using external modulators and demodulators, and no polarization multiplexing.

    7. Re: So much slower than 100GBase-RZ...? by Geordish · · Score: 3, Informative

      Not quite. Those optics use DP-QPSK, which uses mathematical magic to cram 4 bits worth of information into one symbol. This means the optics only need to operate at 25Gbps to supply a 100Gbps line rate.

      DP-QPSK is a whole load of magic I don't understand.

      If DP-QPSK can be used with this technology, it seems to imply 200Gbps optics are not too far away.

    8. Re: So much slower than 100GBase-RZ...? by Geordish · · Score: 1

      Nope, that is 100Gbase-(L/S/E)R4. ZR uses DP-QSPK.

    9. Re: So much slower than 100GBase-RZ...? by Anonymous Coward · · Score: 0

      Yes, but if you can achieve that with a single line and single color, then imagine what you could do if you added more lines and more colors.

      Basically, it's a significant advancement because it will let us use other existing technology to achieve even faster speeds. An example her is to say if they used 3 colors per line and 8 lines. that would be 57Gbps * 8 lines * 3 colors = 1368Gbps or 1.368Tbps, depending on weather color filters interfere with the technology or not. And this can be expanded upon to achieve even faster speeds.

    10. Re: So much slower than 100GBase-RZ...? by ChrisSlicks · · Score: 1

      The article is terrible so we don't really know what was done to achieve the results (other than it was a single fiber). Also QPSK actually only doubles the symbol rate as it sends 2-bits of information per symbol (4 possible permutations).

    11. Re: So much slower than 100GBase-RZ...? by KGIII · · Score: 1

      I love it when you talk dirty. ;-)

      No, seriously, I do. I like it when I have to double click to highlight a word, right click, search, and then figure out which result is most meaningful. I learn new and interesting ways to break things.

      --
      "So long and thanks for all the fish."
    12. Re:So much slower than 100GBase-RZ...? by Lumpy · · Score: 2

      Because 100GBase-ZR is useless over a 1000 mile long single mode fiber.

      --
      Do not look at laser with remaining good eye.
    13. Re: So much slower than 100GBase-RZ...? by Geordish · · Score: 1

      Yes, QPSK does send 2 symbols, but DP-QPSK doubles that to 4 symbols.

    14. Re:So much slower than 100GBase-RZ...? by ffkom · · Score: 1

      What makes you assume the technology described in the linked article is good for more than the 80+ kilometers that 100GBase-RZ achieves?

    15. Re: So much slower than 100GBase-RZ...? by Anonymous Coward · · Score: 0

      We can already do 400Gb/s per super channel and 80 supper channels per fiber and a 1400km range with commercially available DWMD tech. For several years.

    16. Re:So much slower than 100GBase-RZ...? by ChumpusRex2003 · · Score: 1

      100GBase-ZR achieves its 120 Gbps line rate by using a complex modulation scheme to encode 3 bits per symbol. 2 bits are transmitted using QPSK and 1 bit is transmitted by choosing either horizontal or vertical polarization.

      The issue with long-haul transmission is that you only have a limited bandwidth available which works with optical amplifiers and avoids the water "dip". It's common to use DWDM techniques to cram multiple individual streams onto a single fiber. This yields just under 100 usable channels of 50 GHz bandwidth. The advantage of Juniper's proprietary "100GBase-ZR" protocol is that it needs only just over 40 GHz of bandwidth, so fits into a single channel.

      The technology described in the article is about a new type of VCSEL laser diode. These have been widely used as transmission elements, but have traditionally had limited on-off bandwidth, hence early 10 Gbps transceivers often used much more expensive lasers such as Fabry-Perot lasers, and devices operating on 25 Gbps lanes have resorted to the even more expensive Mach-Zender optical modulators. VCSELs are easy to fabricate and cheap, hence if a simple On-Off modulation is sufficient then they are ideal (not to mention that carrier-switched modulation is easy to demodulate, unlike the nightmare which is DP-QPSK). The disdvantage with OOK is that the bandwidth equals the symbol rate, so 50 Gbps OOK would not fit into a single 50 GHz channel with guard bands, and would require a larger channel allocation on a trunk fiber.

  4. Error free? Even more remarkable is (wait for it) by Provocateur · · Score: 1

    that the rooms at either end were at room temperature; just ask renown intellectual Mr Steven Wright.

    --
    WARNING: Smartphones have side effects--most of them undocumented.
  5. Re:Error free? Even more remarkable is (wait for i by U2xhc2hkb3QgU3Vja3M · · Score: 1

    Aren't all rooms at "room temperature"? - Jerry Seinfeld*

    * this sounds like something he'd say.

  6. Record is over a petabit per second by Anonymous Coward · · Score: 0

    Very misleading - this is a record for a particular type of laser but not for fiber optics itself. NEC and Corning exceeded a petabit per second over fiber a few years ago:

    http://optics.org/news/4/1/29

    1. Re: Record is over a petabit per second by Geordish · · Score: 1

      That was using multiple wavelengths on multiple fibres. This appears to be one wavelength on one fibre. Different kettle of fish.

      "Researchers from the NEC Labs in Princeton, NJ, USA, and from Corningâ(TM)s Sullivan Park Research Center in Corning, NY, successfully demonstrated ultra-high speed transmission with a capacity of 1.05 petabit/s (1015 bits per second) over novel multi-core fiber that contains 12 single-mode and two few-mode cores by employing the advanced space division multiplexing scheme and optical multiple-input multiple-output signal processing technique."

    2. Re: Record is over a petabit per second by Anonymous Coward · · Score: 0

      I think you read it wrong - it was multiple cores on the same fiber:

      "the aggregate transmission capacity was 1.050 Petabit/s, which the partners claim is the highest capacity over a single optical fiber reported so far."

      This was also done at over 50km as opposed to the same room in a lab somewhere.

    3. Re: Record is over a petabit per second by Geordish · · Score: 1

      Multiple cores, multiple fibres. Effectively the same thing. I believe this test was over 1 core. That's the big difference.

    4. Re: Record is over a petabit per second by macpacheco · · Score: 1

      Its NEVER single fiber. Its always two strand (two fibers) one TX, one RX. DWDM requires one fiber for each lane.
      Anyhow, nobody lays optical cables with 2 strands for long range networks.
      Its always 12-288 strand cable. And 12 stand is being really cheap. 36 strand is a more common low end.
      So a 36 stand cable allows for 18 DWDM systems, 1 Tbps each, or 18Tbps of bandwidth on a low end cable.
      This is another case of state of the art (regardless of cost) advancing, which someday will trickle down to real world optical links on land networks first, then years later onto oceanic cables (which require very long distances between active regenerations).

  7. Error correction by Anonymous Coward · · Score: 0

    Error-free is nice and all, but if you could transmit substantially faster with a relatively low error rate, wouldn't that be a worthwhile trade?

  8. BT fiber @ 800Gbit nearly 3 years ago by Anonymous Coward · · Score: 1

    https://tech.slashdot.org/story/13/05/24/015240/bt-runs-an-800gbps-channel-on-old-fiber

    1. Re:BT fiber @ 800Gbit nearly 3 years ago by Geordish · · Score: 1

      "The superchannel is an advanced dense wave division multiplexing (DWDM) technique, created by combining multiple coherent optical signals into one channel"

  9. Error free? by rfengr · · Score: 3, Insightful

    Lousy article; no details. There is no such thing as error free, so what is the threshold? 1E-9?

    1. Re:Error free? by Anonymous Coward · · Score: 1, Informative

      Oh, you twat. If a transmitted unit contains an error it is retransmitted, which reduces the overall throughput. This is digital communications, so there is such a thing as error free.

    2. Re:Error free? by Anonymous Coward · · Score: 0

      The same way harddrives are digital and error free. ZFS is pointless.

    3. Re:Error free? by Anonymous Coward · · Score: 0

      Oh please! what is meant is that the actual throughput once errors and subsequent resends are taken into account (i.e. the actual throughput of fixed-for-error data) is 57Gbps.

      Where is the pedantic mod when you need it?

  10. Gimme gimme gimme! by Anonymous Coward · · Score: 0

    I need some more! Gimme gimme gimme! Don't ask what for!

  11. Error Free? by TechyImmigrant · · Score: 1

    For any reasonable bit error rate, it is entirely possible to build a chip that can do error correction at 57 Gbps.

    --
    I should use this sig to advertise my book ISBN-13 : 978-1501515132.
  12. Local speeds by Jason+Levine · · Score: 1

    Researchers achieve 57Gbps through fiber optic lines. In local news, I'm still stuck at 15Mbps because Time Warner Cable is a local monopoly and thus has no incentive to upgrade their speeds.

    --
    My sci-fi novel, Ghost Thief, is now available from Amazon.com.
    1. Re:Local speeds by KGIII · · Score: 1

      At home, I pay for 12 and get about 14. Honestly, that's more than adequate for my needs. :/

      I do wish I could actually throttle it back a little and use some of it for upload as upload is only about 1.5 Mb/sec, averaged out. I'd go down to 10 down for 5 up. I'd still be fine. I host all sorts of things off it. I hang stuff off the network like a Christmas tree - though I do have three disparate lines and everything on the network is actually heavily locked down BUT I am, technically, using it right now.

      Right now, I'm connected via VNC and sending this to you. I've been doing this since September of last year when I hit the road. (It helps to have a static IP but that's not required.) Of course, only a certain subset of IPs are allowed to connect at all - the rest get dumped to the bit bucket - but, damn it, there's all sorts of neat things running. I just don't want *you* to access 'em without permission. ;-)

      As an aside, in the lower-right... Sometime in the past few days, a 1x1 pixel called js.gif has appeared. Doth the tracking pixel of doom rear its ugly head? (It's blocked by default so it's a silly looking failed loading icon - I've not yet masked it with uBlock.) Web beacons haven't died an ugly death yet? *sighs* I'm in the process of building a new site - in fact, a network of 'em (ideally). One of the things I'm trying really hard to do is to keep everything hosted locally. However, it's gonna end up with the almighty Google AdSense on it so I'll probably have that and the damned analytics. They're just too damned handy. However, I'll go so far as to tell everyone how to block 'em. ;-) (It's the principle, not the money - and a long story.)

      --
      "So long and thanks for all the fish."
  13. Fiber optics have a limit by manu0601 · · Score: 3, Insightful

    Summary says:

    breaking previous theories that fiber optics have a limit in how much data they can carry

    In other words, fiber optics would have no limit in how much data they can carry, which it utterly bullshit. How someone could write that?

  14. Re:Error free?Even more remarkable is (wait for it by Anonymous Coward · · Score: 1

    close but no cigar. google steven wright

    captcha "jammed" as in comedy

  15. Awesome by Dunbal · · Score: 1

    Able to reach your monthly data cap in less than 6 seconds!

    --
    Seven puppies were harmed during the making of this post.
    1. Re:Awesome by Anonymous Coward · · Score: 0

      Wait, you have data caps with your fibre?

      lol, how backwards.

  16. No hard drive can write data that fast by Anonymous Coward · · Score: 0

    No hard drive can write data that fast! While it is fast, is it useful?

    1. Re:No hard drive can write data that fast by chebeba · · Score: 0

      Ask (for example) the guys at CERN. The LHC produces data at about 25GB/s while running all experiments, and all this data has to be sent over long distances to be stored and processed.

    2. Re:No hard drive can write data that fast by ledow · · Score: 1

      What idiot would use a hard drive to store and record everything they send? No, this is about processed data.

      P.S. Large storage arrays, especially those tied to high end data systems, can easily manage this.

      You think Amazon aren't pushing Gbps? Google? You think millions of people transferring stuff to their DropBox isn't collectively more than this?

      You really need to think before posting on an IT board.

  17. No, VCSELs do not emit cavity surfaces... by chebeba · · Score: 0

    The name is confusing, maybe.
    Laser diodes usually emit the beam parallel to the chip surface, while the beam of VCSELs is perpendicular to the surface.
    This gives a number of advantages in the production process, increasing yield and the density of diodes on the GaAs wafer.

  18. Fiber is not what's important. by Anonymous Coward · · Score: 0

    This really has nothing to do with fiber. It has to do with a particular type of laser transmitter. You have to love it when reporters completely fail to understand the subject of their article and it gets published anyway.

  19. Re:Error free?Even more remarkable is (wait for it by U2xhc2hkb3QgU3Vja3M · · Score: 1

    Watched a few videos of him. Thanks.

  20. Useless metric - need to know bit rate x distance by Anonymous Coward · · Score: 0

    Ugh, how does this stuff get on to /. ?

    Bit rate is a useless metric.

    The metric is bit rate times distance.

    The theoretical limit (set by Brillouin scattering (four wave mixing)) is on the order of Terahertz * Kilometers.

  21. Program a map to display frequency of data... by ChoosyBeggar · · Score: 1

    "At a hundred million megabytes per second, you begin to make out certain blocks in midtown Manhattan, outlines of hundred-year-old industrial parks ringing the old core of Atlanta...” -William Gibson, Neuromancer