Hacker News Reader

Comment by ortusdux | Hacker News Reader

I wonder how long it will be before we see the first hollow-core fiber subsea cables. They are 50% faster, and tests from the last year or two have seen record low signal losses.

https://www.laserfocusworld.com/fiber-optics/article/1419605...

https://www.ofsoptics.com/wp-content/uploads/Hollow-Core-Fib...

View on HN

I might be totally wrong; It seems likely to me that, due to capillary action, if a hollow undersea fiber gets physically cut then seawater would flow into the hollow center.

The ends of the fiber might be at different depths with a pressure difference that could move water a long way into the fiber. I imagine the length that water got into would be ruined even if the water was pushed out again.

I conjecture that undersea hollow-core might end up being expensive to maintain.

IIRC individual fibers are terminated every Nkm at a repeater. Not that it wouldn't be spendy, but I would also conjecture replacing a segment of fixed length instead of just gluing the ends back together might still be a reasonably strong constraint on unplanned repair cost (and also probably providing a pretty strong lower constraint as well--notably higher than solid core).

I don't think they are terminated every Nkm anymore. We have been able for quite some time to re-energize the signal directly in the optical core without needing to convert it to electric then back.

https://en.wikipedia.org/wiki/Laser_pumping

If a cable is long enough to require amplifiers, they are spliced in. Couplers/circulators need to be in line with the doped section of fiber that forms the laser amplifier, both before and after. One of the couplers injects the pumping beam from another laser into the doped section. This necessitates splicing at every amplifier. Also, almost every strand in the cable would require amplification.

In addition there are feedback, failover and monitoring functions that require more optical components to be included, and it's likely that this type of functionality will increase as demand for improved latency and reliability increases, and new cable networks are built.

If there's a leak that would allow water access to the core, the signal's already gone.

And, a hole that small in a block of glass could withstand a titanic amount of pressure.

Cables like this would have spare strands to be able to handle the situation in which a single strand was compromised by a break or pinhole, and possibly a different type of jacket on each strand to manage the stresses. If a section is badly damaged beyond the point of patchwork repair, then the only hope from that point is that the customers using the strands/wavelengths on the cable have an active alternate path while the whole section is dug up and replaced.

I may be even more wrong, but maybe this leads to a more modular way of building/laying these "cables".

I imagine that you need less of them, because "faster" also means longer range until attenuation kicks in. How much more range/less amplifiers compared to current state of the art I don't know. I'd think it should be possible to lay down a new part the whole length between two amplifiers/laser pumps/couplings/ and "plug" it in by remote controlled submersibles, instead of lifting up the broken ends, and "splicing" it on the ship. Like pre-made Cat-5/7 with RJ-45 plugs, so to speak :-)

With hallow-core, Spread Networks[1] could be back on top again!

1 - https://en.wikipedia.org/wiki/Spread_Networks

The "50% faster" stat is talking about latency, which is not really what subsea cables are about (massive bandwidth).

The more important figure is the attenuation at frequency bands used in telecommunication (C-band and L-band). According to the article, hollow core fibers are still much worse than solid core fibers (0.28 dB/km vs 0.14 dB/km).

> They are 50% faster

No, pretty sure light still travels through them at C. What they can do is carry more data, largely by having lower error rates.

'c' is dependent on the medium. The value of c as it is commonly known, 300 million meters per second, is in vacuum. Light traveling through other media is affected by its index of refraction, in the case of silica fiber, that is approximately 1.5 so radiation propagates much slower through silica than it does a vacuum. Since gases have low refractive indicies already, within a hundred ppm or so of a vacuum, you could essentially round air to 1.

I've heard that high frequency traders are interested in Starlink's planned laser links because they could open up routes that are faster than traditional terrestrial fiber.

I think that if you are a HFT, you probably have a server set up next door.

The suggestion is about trading across multiple exchanges, for example between London and NY. Going via Starlink is potentially quicker than a fiber under the Atlantic.

They will have servers “next door” to the exchanges, but need the servers to have incredible low latency connections to each other.

Yeah, that would make sense. There are links that have been built by various HFT firms and banks [0] [1] that use microwaves instead of fiber buried in the ground simply because of this speed-of-light-in-media limitation. They can shave a few hundred nanoseconds (or something, I don't want to do the math right now) because of a higher signal propagation speed and get a trade in faster than their competitors. Same thing with a laser link like this.

Edit: cf.

[0] https://arstechnica.com/information-technology/2016/11/priva... [1] https://www.reuters.com/article/us-highfrequency-microwave/l...

I know the main people behind much of the hollow core work being done. Much of their financing is coming from HFT related firms.

Light only travels at C in a vacuum (so it doesn't actually travel at C in a standard fiber optic cable, it's actually much slower).

"c" is the speed of light in a vacuum. Traditional fiber optic cables are very much not a vacuum, with an index of refraction of ~1.5, so light travels through them at ~2/3c. In contrast, light actually travels at nearly c through hollow core cables.

It's so nice when someone other than myself is confidently incorrect.

Reply on news.ycombinator.com