How fast is that? Download one gigabyte in 0.2 milliseconds! Less than the blink of an eye, literally.
"... researchers at the Technical University of Denmark have now reclaimed the record for the fastest network ... using single multi-core optical fibre, which was developed by Japanese firm Nippon Telegraph and Telephone Corporation (NNT)"
"A research group at the Technical University of Denmark (DTU), which was the first to break the one-terabit barrier in 2009, has today managed to squeeze 43 terabits per second over a single optical fiber with just one laser transmitter. In a more user-friendly unit, 43Tbps is equivalent to a transfer rate of around 5.4 terabytes per second—or 5,375 gigabytes to be exact. Yes, if you had your hands on DTU's new fiber-optic network, you could transfer the entire contents of your 1TB hard drive in a fifth of a second—or, to put it another way, a 1GB DVD rip in 0.2 milliseconds."
This type of fibre contains seven cores—glass threads—instead of the single core used in standard fibres, which makes it possible to transfer more data.
Surpisingly, the fibre is the same width as a standard fibre.
SOURCE Because everyone loves The Daily Mail ...
BETTER SOURCE Or, maybe not
(Score: 2, Informative) by ThG on Tuesday August 05 2014, @10:00AM
Seriously, how fast is the fastest hard drive? 500MB/sec writing speed?
That's like, um.. about 10750 slower (assuming mathematical GB of 1.000MB) than what those guys have put through...
Perhaps they should just work on faster hard drives, or how to get more than one bit (light?) stream through such a cable.
(Score: 2) by wonkey_monkey on Tuesday August 05 2014, @10:21AM
Eh? If you can get one bitstream down a cable, you can get arbitrarily many [wikipedia.org] bitstreams down a cable. Not at the original speed, of course, but that's probably not going to be an issue with this cable for a while.
systemd is Roko's Basilisk
(Score: 0) by Anonymous Coward on Tuesday August 05 2014, @10:51AM
You can set up the hard drives in parallel. Then you only need 10750 (under certain assumptions) hard drives with a speed of 500 MB/sec each to cope with the link speed.
(Score: 4, Funny) by acid andy on Tuesday August 05 2014, @12:07PM
In that case, the logical solution must be to replace all disk drives with loops of optical fibre. All your data just revolves round and round and round the optical magic roundabout until you need to access it!
If a cat has kittens, does a rat have rittens, a bat bittens and a mat mittens?
(Score: 1) by Kunasou on Wednesday August 06 2014, @08:50AM
Cool, want to see one of those working.
But this speeds are only for ISP links.
(Score: 4, Informative) by isostatic on Tuesday August 05 2014, @12:19PM
I can't believe this has been modded up, typical home users thinking of their bandwidth from their pc to their cheeto machine.
Typical bitrates through undersea cables exceed 500gbit per fibre, and the need isn't going down any time soon. WACS I believe can (if lit) take well over 1Tbit per fibre, for a total throughput of 5Tbit.
With 2tbit of bandwidth available, I could rig a 20 camera 8K setup in Rio for the 2016 Olympics and bring back the material back to my production centre, saving a hella-lot of money. Those bits would never see a hard drive, but they'll see 8000 miles of undersea cable.
(Score: 2) by LoRdTAW on Tuesday August 05 2014, @01:09PM
Perhaps they should just work on faster hard drives, ...
These speeds would never go anywhere near a hard disk. Ever. They will only see stupid fast ASICS, optics and other high speed interconnects in core routers of tier 1 ISP's. Not the router that serves you porn at up to* 20Mbps.
... or how to get more than one bit (light?) stream through such a cable.
Been around a long time: https://en.wikipedia.org/wiki/Wavelength-division_multiplexing [wikipedia.org]
(Score: 2) by davester666 on Tuesday August 05 2014, @04:42PM
That's just the speed of the connection to Verizon's network. You still won't be able to stream from Netflix without "buffering".
(Score: 0) by Anonymous Coward on Friday September 12 2014, @04:53PM
1r1n9P sjndqqdjtiql [sjndqqdjtiql.com], [url=http://txtpfdncxcbe.com/]txtpfdncxcbe[/url], [link=http://nwugwoljovzf.com/]nwugwoljovzf[/link], http://yfkmuatfcnkv.com/ [yfkmuatfcnkv.com]
(Score: 2) by wonkey_monkey on Tuesday August 05 2014, @10:11AM
I'm not surprised. In fact, I think I might just die of not surprised.
New things are always getting smaller than the previous things.
systemd is Roko's Basilisk
(Score: 0) by Anonymous Coward on Tuesday August 05 2014, @10:58AM
That's what she said!
(Score: 2) by bob_super on Tuesday August 05 2014, @03:50PM
Actually, the single-mode fiber core is pretty much of fixed diameter (unless they did something here, didn't RTFA). But you can put a few 5 micron cores inside your usual 125 micron cladding if you're very careful about it, and a few of those inside a 900 micron (or 1.25mm) standard-looking fiber.
(Score: 2) by bob_super on Tuesday August 05 2014, @03:52PM
Ouch, I meant 9 micron core.
(Score: 2) by wonkey_monkey on Tuesday August 05 2014, @05:57PM
Oh. That makes it even less surprising, really.
systemd is Roko's Basilisk
(Score: 5, Informative) by sudo rm -rf on Tuesday August 05 2014, @10:57AM
(Score: 4, Funny) by SuperCharlie on Tuesday August 05 2014, @12:21PM
It was then throttled to 200kbs when they tried to view Netflix.
(Score: 2) by present_arms on Tuesday August 05 2014, @01:18PM
That's what happens when you get Comcast any where near it :P.
http://trinity.mypclinuxos.com/
(Score: 2) by Leebert on Tuesday August 05 2014, @01:57PM
This sounds fairly tricky to work with. Having done my fair share of work with standard optical fiber, it's already enough of a pain to terminate and splice with just a single core where everything is perfectly symmetrical. I don't want to imagine how much of a pain it would be to line up 7 cores. I have little doubt that this will require very expensive specialized equipment, and won't make it anywhere near the level of accessibility of standard OM2/OM3 fiber.
(Score: 2) by bob_super on Tuesday August 05 2014, @03:45PM
In multi-fiber cables, you just separate the fibers a few cm before you terminate each into the connector(s).
(Score: 2) by Leebert on Tuesday August 05 2014, @03:55PM
This isn't multi-fiber, it's multi-core. Unless someone is severely butchering fiber terminology...
(Score: 2) by MrGuy on Tuesday August 05 2014, @03:14PM
I get multiple emitters/multiple paths/same bundle offering speed bursts.
What I don't get (and what both the referenced articles don't refer to) is how multiple cores HELPS with a single emitter. If we have a single emitter, we have a single wavelength to work with, and a single "flashing light" to transmit at that wavelength.
I had thought (someone can correct me) that current single-core fibers were already close to theoretically maxing out how much information a single pulsing laser could transmit information (i.e. we can't pulse the laser on and off more quickly without the pulses interfering with each other.
So how does having more cores help? I see some speculation on multiplexing different frequencies with different messages, but that seems to require multiple emitters.
What's the theoretical principle that multi-core/single-emitter could be exploiting to be faster?
(Score: 0) by Anonymous Coward on Tuesday August 05 2014, @05:30PM
dunno. maybe with one tunnel you hit a maxspeed X after which the errors are to much to correct.
so maybe if you use multiple tunnels and shot the identical stuff thru it you can improve the error-correction "somehow" and thus up the speed until you again hit a "error-correction" barrier?
(Score: 0) by Anonymous Coward on Tuesday August 05 2014, @04:41PM