[Freedombox-discuss] We do need mesh networking

[Michael Blizek]

Hi!

On 10:04 Thu 24 Feb     , Luca Dionisi wrote:
...
> You should distinguish mesh networks and ad-hoc wireless networks.
> Mesh networks are not implicitly wireless, nor ad-hoc.
> The problems in that paper refer to ad-hoc.
> 
> Make a mesh network variegated and you won't have those problems.
> e.g. Node A, B and C create an ad-hoc network at channel 1. Node D
> acts as an access point for stations node E, F and G at channel 11.
> Node A, E and J are connected in a wired LAN. Node J also forms
> another ad-hoc network, say at channel 1 because it is not in the
> range of A, B and C.
> The more the network is variegated, the less you face problems that
> are related to ad-hoc wireless networks.
> Perhaps, it will also bring new sort of problems, admittedly.

Using multichannel radio, wired connections and other technologies will help
you increasing line speed and limit interference. Actually it might help a
*lot* in terms of performance. However one scaleabilty factor remains: The
amount of data each individual node has to forward grows linear with the
average path length. Given constant available bandwidth, you can choose
between "short path and high bandwidth" or "long path and low bandwidth". You
can get around this by adding more bandwidth, but...

	-Michi
-- 
programing a layer 3+4 network protocol for mesh networks
see http://michaelblizek.twilightparadox.com