Awesome!
Amazing work!
UK boffins DOUBLE distance of fiber data: London to New York WITHOUT a repeater
Researchers at University College London say they have found a way to dramatically increase the distance data can be sent intact over ordinary fiber optic cables – by partially working around the Kerr Effect. A phenomenon discovered by Scottish physicist John Kerr in 1875 limits the distance coherent data can travel. The …
-
-
Friday 6th February 2015 22:13 GMT Anonymous Coward
How many homes are 3,500 miles from the nearest exchange?
The challenge in last mile networks isn't technical, it's commercial. How do you justify spending £2k+ per installation for a service people want to pay £25 a month for? It's a problem that's not unique to BT - no telco has solved it yet.
-
-
-
Thursday 5th February 2015 09:37 GMT Lee D
Re: Testing
When you buy fibre, does it come as an 80km long parcel?
No, it comes wrapped up in a coil.
Put coil on floor. Grab both ends. Plug into your equipment half-inch away from each other. Voila! Test circuit.
Hence why a lot of these kinds of things don't work (or are severely reduced) once they are deployed in the real world.
That said, it's an interesting technique and if you think of the amount of SFP in the average business, getting the same speed over twice the distance may eventually equate to twice the speed over the same distance. This is the sort of stuff that ends up being in 100Gb protocols and you never know about it.
-
Thursday 5th February 2015 10:22 GMT Cliff
Re: Testing
About time someone taught that showoff Kerr a lesson.
Actually the opposite - fuck me the Victorians were good - it's taken 140 years to find a workaround for a physics problem he spotted before my great great grandfather was born. I'll bet Kerr could have programmed the video.
-
Thursday 5th February 2015 17:23 GMT BreeBree
Re: Testing
Having worked for a Very large Fiber Optic Provider. Yes it does come in 80Km and more lengths.. to use USA measurements, the spools to be installed for the contract are hooked up together and yes it goes thru the WHOLE thing, to test all the fiber and find out if any is bad, so all that would happen in deployment would happen in the factory even when testing 3500MILES!! of single mode cable running 256 channels at 40+gig/sec per channel with NO repeaters. This has been done and are installed in the world now so this is old news that it can be done.
Now this might be a better way or new to the world, but going across the pond is old hat for some companies that use NO repeaters, I know this was being done 8+year ago, as I was on the team that tested it and help deploy it.
Note for Fiber its not like copper wire , it doesn't radiate EMF like a coil or wire does when voltage/current flows. the problem is noise and attenuation of the light itself from impurities. Better cable now and better head end electronics have made repeaters a non issue. so even in lab/factory if you put one station next to another and run it thru all the fiber its just like being out in the installation.
-
Friday 6th February 2015 22:22 GMT Anonymous Coward
Re: Testing
"Put coil on floor. Grab both ends. Plug into your equipment half-inch away from each other. Voila! Test circuit.
Hence why a lot of these kinds of things don't work (or are severely reduced) once they are deployed in the real world."
Only for PoC testing. No-one's going to deploy stuff like this without field testing. It's pretty expensive to wind it all back if it doesn't work. There's plenty of unlit fibre around the world to test stuff properly.
-
-
-
Thursday 5th February 2015 10:02 GMT Ragequit
Cool
Cool. Now if they could only do the reverse. That is reduce the effective range but increase the bandwidth/channels. Then maybe cable companies would offer higher speeds... oh wait.. no they'd probably just put twice as many households on a single fiber run. Nvm.
Seriously though I understand how having to use repeaters for anything can be a pain for a number of reasons. I imagine especially so for underwater cables.
-
Friday 6th February 2015 22:37 GMT Anonymous Coward
Re: Cool
"That is reduce the effective range but increase the bandwidth/channels"
That's a trivial engineering problem. You don't even need to reduce the effective transmission distance.
What's not trivial is the commercial problem of justifying the capital outlay to upgrade an existing network in a market where price is declining.
-
-
Thursday 5th February 2015 15:22 GMT Anonymous Coward
Establishment he say "Inappropriate use of taxpayer money"
A government spokesperson said today "If there aren't any repeaters, how can our trustworthy security forces tap the fibres for our safety? It's this type of so-called research that aids terrorists, child-snatchers, and ICDJs*
Oh, I wish that wasn't so plausible a response.
* - Invented Crime Du Jour
-
Thursday 5th February 2015 17:51 GMT Dave Bell
That awkward memory moment.
It wasn't until somebody mentioned the Septics that I recalled 80km was 50 miles, near enough.
In my distant youth I measured a field in Chains and Links.
And I can remember MAFF, as it was then, specifying ridiculous precision for land measurement when they introduced a new EU scheme.
Now I can get a laser rangefinder for under a hundred quid. I hope somebody remembers the difference between accuracy and precision: I am not sure the guy writing the Amazon description knows that.
-
-
Friday 6th February 2015 02:05 GMT Alan Brown
Re: This would be big if it can also be applied to copper
If it could double the distance then it might make g.fast almost practical, but they're hard up against the shannon limits so I doubt they'll get much improvement.
cpe>
Extended Port Status
=================
Bme: 1 Port: 1
Downstream line rate: 85760 kbps
Upstream line rate: 21600 kbps
Downstream attainable payload rate: 91948 kbps
Downstream attainable line rate: 103628 kbps
[snippage]
Tx total power -8.0 dbm
FE Tx total power 12.6 dbm
VDSL Estimated Loop Length : 536 ft
G.Hs Estimated Near End Loop Length : 4495 ft
G.Hs Estimated Far End Loop Length :0 ft
I can almost spit on the cabinet - it's about 30 metres away, but across the road.
Having 160 metres of cable run isn't at all unusual for a short line-of-sight to the cabinet, but bear in mind that some houses may end up with 500 metre runs and G.fast will be lucky if it can get to 100m in real-world installs without speed falling below 100MB/s
-
-
Friday 6th February 2015 22:46 GMT Anonymous Coward
Re: This would be big if it can also be applied to copper
"this technique reduces the signal distortion which would reduce the noise, effectively raising the SNR."
SNR remains the same. What changes is the threshold for effective signal discrimination. This doesn't solve the crosstalk problem however, so I'm not sure this can be applied effectively to a typical copper distriution network. It solves the wrong problem.
-
-
-
Friday 6th February 2015 22:31 GMT Anonymous Coward
Re: This would be big if it can also be applied to copper
The issue with copper is less about coherence and more about crosstalk. This development is useful for a single bearer considered in isolation - which fibre is due to the medium - but with multiple unshielded bi-directional copper bearers the problem is different.
-
-
Friday 6th February 2015 07:01 GMT John Smith 19
Note the part about *existing* fibre, not tricky stuff with FO ampliers.
There's a lot of that in the ground/ocean already and replacing a bunch of repeaters with basically a short length of passive fibre (if you can't pull one of them through into the chamber and join them directly) is going to save some people a very great deal of money.
-
Sunday 8th February 2015 23:09 GMT Jan 0
Kerr
I have a vague memory of using* or maybe seeing a demonstration of the use of Kerr Cells to measure the speed of light when Iwas a sixth former. (So long ago that it wasn't clear that the 'speed of light' was a constant regardless of frequency.)
*Well, we did alll kinds of things in school science labs back then that today's students will only see as a video, if they're lucky.
Biting the hand that feeds IT
