Kwality IT reporting...

[Johnny B Good]

That's basically the method though the 20 to 30 volts drop might have been shared to repeaters in clusters of 3 or 4 at a time. Modern kit using regenerator/laser amplifiers may be able to manage on less voltage, plus the repeater spacing on fibre links can be an order or two greater than the 2Km spacing that was once used on undersea co-axial cables using analogue RF amplifiers back in the days of FDM. The shorter undersea fibre optic routes of 100Km or so can manage without resorting to undersea repeaters.

On longer, trans-oceanic routes measuring several thousand kilometres, each end of the cable would be fed with an HT DC voltage measuring several Kilovolts! (7KV at each end of the TAT cables being laid down in the late 70s / early 80s afair - positive one end and negative the other to provide a total of 14Kv over co-axial cable that didn't need to be rated for the full 14Kv).

[Jacky Chance]

Pity about what powers the repeaters.

[Dave Liquorice]

They might be using light for that as well. I know it's been researched and they could get a useable amount of power at the reciving end. Don't think it was very effcient though but things could have improved.

[Jacky Chance]

No they don't on that particular cable.

I know it's been

I don't believe that with a cable anything like that long with the number of repeaters that cable uses.

Don't think it was very effcient though but things

No they have not. There is nothing like enough light to do that.

[Henry Law]

Every so often I see a really fascinating USENET post which reminds me why I spend so much time on t'interweb. This was one of them; thank you.

[Paul]

A 4800KM fiber might only need ten repeaters. But that won't change the nature of the cable bundle, until the repeaters are removed entirely. And that could happen. There are people working on DSP techniques. I was unable to find any practical (field) examples of that.

Because it's a winner-take-all field, you're not likely to find the very latest developments, until they're on a ship out to sea.

Paul

[Mike Tomlinson]

En el artículo , Henry Law escribió:

Agreed. Thanks Johnny.

[Mike Tomlinson]

En el artículo , Dave Liquorice escribió:

Only small amounts, and very probably not enough for undersea repeaters.

Interesting research though.

[GB]

Apart from which, surely the light travelling down the optical fibre

*is* an electromagnetic signal?

How it would get out is another matter.

[Tim Streater]

Indeed it is, but if the light in the fibre core is escaping to an extent that the sharks notice it, the operator of the cable would be in serious trouble.

[dennis]

Soliton optical solutions may be the answer, they can use optical amplifiers rather than repeaters. There have been experiments with thousands of killometres of fibre AFAIK.

[The Natural Philosopher]

That sounds about right. I worked* on P-TAT repeater software - shark attack was a standing joke.

Fortunately it only affected shallow bits of cable

*or I would have done if anybody had been able to explain what they were trying to do.

[The Natural Philosopher]

Not with an 8086 involved.

The repeaters weren't just dumb amps,. They had the ability to switch between fibres and IIRC report back on faults

And if there had been enough light power to power an amp, they wouldn't have needed one! Doh!

[The Natural Philosopher]

Its the magnetic field that the sharks detect IIRC. You wont get a static magnetic field with light, and the whole point of the fiber is to make sure it stay trapped inside

[Johnny B Good]

Well, thank you, Mike and Henry. These longer posts usually attract bric bats (?) rather than accolades. :-)

I suppose the shorter than usual length of this post when I find myself drawn into "Essay Mode" must have tempered the (in retrospect, often understandable) urge to fling bric bats in my direction this time.

Damn it! Here I go again, turning a simple 'thank you' into a mini essay! :-(

[John Rumm]

Being a fan of the "fuller" answer myself, I would add my thanks also ;-)

[pamela]

You wrote you worked on P-TAT ...... and also that you would have worked on P-TAT but didn't.

Which is it?

[The Natural Philosopher]

Well I spent 6 months bashing a keyboard, looking out of the window and I even brought a telephoto to photo people in the park outside, but as far as actually generating anything of use or value, I have no idea.

Since no one seemed to know what any of the program specifications meant.

A couple of blokes and I reckoned we could have written the whole thing from scratch if we had thrown away the 'z notation' sacked the management and RMX 86, written a lightweight multitasking kernel, and hacked the thing together, in about 3 months

Instead it had about 35 programmers working on it.

And they still were when I left.

Money was OK tho.

[pamela]

You wrote, "I even brought a telephoto to photo people in the park outside". Well that's seems unusual!

[Paul]

In the Real World, things are decidedly utilitarian. There are a *lot* of repeaters on this one. I guess like most things in life, the emphasis is on quantity over quality. And this cable only cost $300 million. It's a mere 2cm in diameter (and is buried). Not a lot of diameter for reinforcement. So a shark who wants to bite this one, will have to catch the cable as it goes over some "bumps".

And this describes a state-of-the-art coherent detection (DSP) method at 100Gbit/sec. The repeater distance is consistent with the info in the previous article.

And I get a third (mis-stated) total bandwidth figure in this video. I think in journalism school they emphasize "just put in some numbers and let God sort it out". Maybe the total cable is rated 60Tbit/sec (6 pair times

100Gbit/sec times 100 wavelengths) ? I can see a power converter in the middle of the unit, but cannot recognize much more of what is in there. The interesting bit isn't in the transparent section. The repeater looks kinda small for digital repeatering, so maybe an optical technique is used.

Apparently there are 6 fiber pairs in the cable, and for its part in the investment, Google gets one pair for their own usage (so they bought a sixth of the cable).

And that thing was supposed to be turned on, a couple days ago. (Now they have to call tech support, turn it off and on again, set their browser to 192.168.1.100 and so on.)

Paul