Getting A Mesh Network To Work In The Real World Takes Some Work
from the not-so-easy dept
People have been talking about mesh networking for a while, and there have been plenty of laboratory demonstrations that show how it works, but when brought out into the real world, all sorts of problems occur. Now, those problems are finally starting to get addressed. While experiments are being done on real world mesh networks all over the place, here’s an interesting report looking at a specific mesh network call “Roofnet” being put together and tested by students at MIT. They’re using WiFi, of course, but have discovered all sorts of interesting challenges that no one predicted. Nodes that are closer to each other in terms of distances aren’t always the best ones to communicate with each other and different WiFi equipment (both access points and cards) seem to vary drastically in how far and how fast they can send a signal (even when advertised to have the same range). Furthermore, the network is constantly changing, and sometimes you’ll be able to send data one way, but not back via the same path. Also, the typical “ping” used to judge whether or not a node can find another node is often inadequate in determining whether or not the two nodes can actually send any real data back and forth. They’re working on all of these issues at MIT, which could help solve the “last mile” problem of high speed internet access – and also make it easier for “mobile hotspots” to work. If they can figure out how to get mesh networks to work, then imagine having a WiFi access point in every car, and being able to have the signal bounce from moving car to moving car until it finds a fixed connection.