Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

> Over long distances, fibre optic would have lower latency so it'd be shorter if taking the same path today.

Source that claim, it's well understood the speed of light is around 66% due to refractive index in glass.

It gets weird with telegraph cables and capacitance, wikipedia at least touches on it here: https://en.wikipedia.org/wiki/Speed_of_electricity





I should have definitely qualified that statement. Technically, electrical signals over copper are "slowed down" less than light through fibre optic cables. However there's attenuation, electromagnetic interference, and other signal loss for electrical signals that (for long haul cables) will mean you will need repeaters that add significant amounts of latency. On top of that, the higher you try and up the frequencies, the worse these problems get.

For some medium-haul stuff, it wouldn't surprise me if you saw copper still being used for lower latency (eg between datacenter sites for flash-trading), but otherwise it's just not economical.


That's the point of hollow-core fibre which is absolutely being used where decreasing latency even by small amounts is worth it.

> That's the point of hollow-core fibre

Ok, how does that work though? I understand the concept of lower attenuation since air/vacuum has less molecules to get in the way. Less repeaters, should have less system latency.

What I don't understand is how light is moving through what is a hollow bendable medium. Is the tube that it's in reflective and there's just less time it's passing through it? I guess that's the main one in commercial use to shave some time off, reading about it here: https://en.wikipedia.org/wiki/Photonic-crystal_fiber


It works in that light will travel faster in a less dense medium. Remove the relatively denser glass for gas/vacuum.

Also the way fibre works is commonly misunderstood. The light isn't bouncing.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: