WDS & spanning tree query
ms at ecs.soton.ac.uk
Wed Nov 6 21:08:03 EST 2002
I have a few questions about WDS in HostAP... I couldn't find any
discussion of this in the archives, but my apologies if it has been
HostAP uses spanning tree it's WDS, correct? I might have completely
mis-understood here, but since spanning tree is designed to *remove*
redundant links in a network isnt this a bad choice in the context of
a wireless mesh?
To illustrate what I mean, consider the following 5 node mesh, where all
communication is done over wireless. Nodes (1,2,3), (2,4) and (3,4,5)
are within range of each other, so ideally the topology would look like
this, with the maximum number of connections.
\ / |
However if you ran spanning tree on this, then one outcome might be a
topology which looked more like this:
So now for 5 to communicate with 4 the traffic goes through every
other node in the cell, despite the fact that they are within range of
one-another and so could in theory communicate directly?
Is my understanding of this correct and if so is it possible to fix this
I suppose bridging is an easy "off the shelf" solution, but would it be
possible to make something better? :) I've been giving this some thought,
and I think it would be plausible to have a fully connected network so
long as every AP knows the topology of the mesh and sends a broadcast
message to all other APs whenever a client becomes associated with it.
Consequently every AP is aware of the location of every client in the mesh
and so can pick the correct node & shortest path on a per packet basis...
Does this sound at all sane?!?
I beleive this is roughly akin to what happens with the normal
distribution service i.e. when multiple APs are connected via wire...
Any info on this would be greatly appreciated :)
Southampton Open Wireless Network (UK)
More information about the HostAP