13 ms·

I did some research on network coding a few years ago (see [1] if interested), and I was very excited by the potential until I read the paper by Li, Li, and Lau

I did some research on network coding a few years ago (see [1] if interested), and I was very excited by the potential until I read the paper by Li, Li, and Lau that proves the maximum gain in throughput we can expect in undirected networks is 2x. https://iqua.ece.toronto.edu/papers/zp-transit.pdf https://iqua.ece.toronto.edu/papers/zp-transit.pdf This makes me think there is no point in introducing all this complexity... just stick to packet networks and routing, and you're at most 2x less efficient than then optimal netwrok coding scheme... [1] A short literature review paper on network coding dating back to 2009: https://minireference.com/static/tmp/network_coding_review.pdf https://minireference.com/static/tmp/network_coding_review.p...

nullc 6y ago
When you've made all other optimizations and your performance is bottlenecked by that last bit of efficiency, then perhaps it is interesting! 2x is nothing to sneeze at. If your application is costing users two million dollars a year in bandwidth ... then halving it would potentially save a million dollars a year-- and could well be worth an impressive engineering effort.
- stelfer 6y ago
  Real world incremental complexity is very high. To say nothing about the lack of an ecosystem. I fondly recall interviewing at one of the faang companies years ago, when they ran out of things to ask me, and the guy turned to me and said: "Let's talk about network coding". My response, "Why would anyone want to do that?", probably didn't help land me the job.