Networking at the Speed of Light: Understanding Fiber Optics (2019)(sikich.com)
sikich.com
Networking at the Speed of Light: Understanding Fiber Optics (2019)
https://www.sikich.com/insight/networking-at-the-speed-of-light-understanding-fiber-optics/
3 comments
I'd like to add that Steenbergen knows networking from more than just the OSI layer 1 part of the optics - he's one of the relatively few people in the ISP backbone industry who knows the optical part of it and also the routing, BGP, router side at OSI layer 3.
There's more than a few router guys out there who mentally abstract away the inter-city links and everything underlying that forms the transport circuit for a router-to-router BGP adjacency. At big ISPs this is handled by a separate team responsible for transport systems, DWDM and metro ethernet circuits.
There's more than a few router guys out there who mentally abstract away the inter-city links and everything underlying that forms the transport circuit for a router-to-router BGP adjacency. At big ISPs this is handled by a separate team responsible for transport systems, DWDM and metro ethernet circuits.
A good, and succinct, little introduction to fibre cables. I didn't realize the speed of light drops 31% inside silicon glass. Thanks for posting.
The velocity factor in different types of copper cables is also around 66% and up to 80%, so can be slightly faster than fiber. Not that it matters much on those distances.
Air dielectric coaxial can get above 90%. RF rides the dielectric in between the shielding and inner conductor, it doesn’t travel inside the inner conductor.
Hence the usage of hollow (air) core fiber to squeeze those extra nanoseconds
This is technical, but the speed of light is constant. This is important to point out because people often get confused with different wave propagation speeds.
It's come up more often recently because it makes for another interesting competitive aspect to LEO satellite internet systems like Starlink. Once they get inter-sat links working, the higher speed of light from paths through air/vacuum would actually give them lower latency over long distances (sufficient to make up for the extra round trip to orbit). It's also the reason high speed traders in some places have gone to the trouble and expense of setting up point-to-point wireless links over long distances; they too can actually beat a regular fiber signal. Photonic band gap fiber is another expensive but interesting tech that might someday be worth it at least for long distance backbones. It's cool we've come far enough to be able to start looking for further improvements like that. For real time international communications, it might make for a real improvement in many applications.
One interesting way that fibers are not immune to RF noise is in the presence of lightning strikes [1]. The strong electromagnetic fields can cause a polarization rotation inside the fiber, which the DSP at the receiving end of coherent links has to track.
[1] PDF Warning: https://www.ofcconference.org/getattachment/d0ec1565-ce81-48...
[1] PDF Warning: https://www.ofcconference.org/getattachment/d0ec1565-ce81-48...
* https://www.youtube.com/watch?v=nKeZaNwPKPo
He covers everything from short-distance intra-data centre to long-distance pan-continental/oceanic technologies.