Slashdot Mirror


Transpacific Unity Fiber Optic Cable Leaves Japan

JoshuaInNippon writes "The 10,000 km (6,200 mile) long Unity fiber optic cable, funded by Google and five East Asian communication companies, left Japanese shores on November 1st to be laid along the northern Pacific Ocean floor. The Japanese end of the cable is expected to be fused to the American end sometime around November 11th. The cable, which was announced in February of 2008 at a cost of around $300 million USD, has the theoretical capacity of 7.68 Tbps, but will be set at a capacity of about 4.8 Tbps (supposedly equivalent to about 75 million simultaneous phone calls) during its initial use. When Unity begins full operation sometime early next year, it is projected to increase internet traffic capacity between the two regions by over 20%, a wonderful boost to transpacific relations!"

10 of 136 comments (clear)

  1. How does that work, exactly? by wandazulu · · Score: 4, Interesting

    So I've got a bunch of cable laying around, figure I'll run my own line from Japan to California. How does that work, exactly? I assume the cable is protected in some extremely strong waterproof and snag-proof sheath, but do they really just roll it off the ship, let it fall to the ocean floor, and there it sits? Do they have to occasionally throw a repeater overboard as well? I've always been curious how we're actually able to have these outrageously long cables under the sea and that it works, and works well enough that I believe cables are still the preferred method of data movement, with satellites being a distant second.

    1. Re:How does that work, exactly? by John+Hasler · · Score: 3, Interesting

      I believe that the cable is plowed in close to shore where possible to protect it against nets, anchors, etc.

      --
      Warning: this article may contain humor, sarcasm, parody, and perhaps even irony. Read at your own risk.
    2. Re:How does that work, exactly? by TBoon · · Score: 3, Interesting

      with satellites being a distant second.

      Have you ever used satellite internet/phones? I have at sea. And disregarding the much lower speeds, the lag makes it highly unsuitable for some usages. We had VOIP phones on our connection. With geostationary satellites the signal take about 200ms just to get from your local point on earth and back down to the other ground-based point. That's very noticable when talking with someone on the phone. Especially when adding a bit more delays at the VOIP-stage and PSTN side too... On the other hand, you can get to pretty much anywhere on the planet within 50ms with a cable. (In theory, disregarding delay at routing, and non-direct routes.)

    3. Re:How does that work, exactly? by sponga · · Score: 4, Interesting

      Yeah screw the article, here is a video and they speak a thousand words. Very cool to actually see the cable being pulled out and what the repeater looks like.

      http://www.youtube.com/watch?v=dOyKdJWPlZY

      SEACOM
      http://www.youtube.com/watch?v=LgKezSWuAGE&feature=related

      Construction of East Africa's undersea fibre optics cable
      http://www.youtube.com/watch?v=dW0Fp-bbKWI

      Alaska Communications Systems Undersea Fiber Optic Projects
      http://www.youtube.com/watch?v=YJt0sh1d-H0

    4. Re:How does that work, exactly? by michael_cain · · Score: 5, Interesting

      Don't forget sharks, that seem to be fooled by the electric field that results from the DC current powering the repeaters, and occasionally attack the cables. I believe newer cables include upgraded armor that is more shark-resistant.

  2. Surprisingly small sounding numbers by Ginger+Unicorn · · Score: 2, Interesting

    For some reason 4.8 terabits/s doesn't sound like that much to me. Obviously it must be since it's boosting traffic by 20% but intuitively I would have imagined another 2 or 3 orders of magnitude for an inter-continental link.

    --
    (1.21 gigawatts) / (88 miles per hour) = 30 757 874 newtons
    1. Re:Surprisingly small sounding numbers by Idiomatick · · Score: 2, Interesting

      The 20% figure is based on the 7.5tb/s speed not the initial 4.8 but still. The % value is the important one to your or I anyways, the actual tb/s figure is meaningless aside from getting a nerd hard on for the bandwidth.

  3. How much of that cost is the cable? by TheSunborn · · Score: 2, Interesting

    How much of that cost is the cable?

  4. Can't wait by raju1kabir · · Score: 2, Interesting

    Looking forward to this here in Malaysia. Global Transit's HQ is just 200m from my house. When I see the truck pulling the final bit of cable wet and dripping from its long sea voyage, I'll slip the dudes a few bucks to tap a slice off for me.

    Seriously, though, this is a country where almost all content of interest is foreign: unlike Japan or Thailand, say, there's no significant local-language content industry. Everyone reads English and/or Chinese and therefore skips straight past the homegrown small-potatoes sites, on to the major international sites (in fact I think most Americans would be surprised how well-integrated Malaysians are into the American view of the web). Every little bit of overseas capacity makes a big difference. Most Malaysian users' home broadband is capped to a measly maximum 4mbps because demand for bandwidth so far outstrips supply.

    --
    "Patriotism is your conviction that this country is superior to all other countries because you were born in it." -- GBS
  5. 75 million? by adenied · · Score: 2, Interesting

    The article submitter seems skeptical of 4.8 Tbps being 75 million simultaneous calls.

    So is 4.8 Tbps really 75 million simultaneous phone calls? Let's do some simple calculations. If we want to go with exactly 4.8 Tbps we can say that's 480 OC-192 circuits. An OC-192 is equivalent to 192 DS3s. So that gives us 92,160 DS3s. Each DS3 carries 28 T1s. So that's 2,580,480 T1 circuits. Ignoring signaling channels and going with a standard DS0 signal of 64 kbps you have 24 channels per T1. Uh oh, that only gets us 61,931,520 voice circuits.

    So where do we get 75 million from? Bad math actually, at least as far as any telecom geek is concerned. If you take 4,800,000,000,000 bps and divide that by 64,000 bps you get exactly 75,000,000. This is very simplified though no matter what the technology being used is. It ignores any overhead in framing and other signaling. Be it traditional telecom circuits like DS3s or packet type networks, you're always going to have overhead. You also need signaling channels to control your voice traffic (unless you want to be old school and use in-band MF or DTMF or something, but I digress). If that's SIP or SS7 or Q.931 ISDN D-channels, you're still taking up space with it.

    I guess all this says is what most people on Slashdot probably already know. Bandwidth is just a number. What you can do with it is an entirely different question.