[MLB-WIRELESS] MeshBox

[paul van den bergen]

On Tue, 2 Dec 2003 10:56 pm, Clae wrote:
> >I don't believe a true "Mesh" network is feasable.  The noise floor
> > created by the mesh increases to the point where the network is useless
> > because it's so slow.  The network effectively kills itself.
>
> I've read a few times that a small number of randomly arranged long
> range links can drastically reduce this problem.
>
> Clae.

pure mesh is doomed to fail for reasons just outlined.  One can extend the 
life by reducing power/range/throughput to just reach the next neighbour at 
the cost of increased hops to a given desination. One can extend further 
through use of highly directional antennas.  

Clearly Meshes are not a viable network option for real time data.  other data 
however - best effort delivery and low b/w is just fine...  off the top of my 
head, proxy squid distribution...

short cut meshes are much more viable.  Small world theory and all that.  the 
difference between a crosslinked heirachical and a shortcut mesh is arbitrary 
though... so start with both and work to the middle.

what am I talking about? imagine APANA merging with MW....

that would really define the last mile problem.


-- 
Dr Paul van den Bergen
Centre for Advanced Internet Architectures
caia.swin.edu.au
pvandenbergen at swin.edu.au
IM:bulwynkl2002
"And some run up hill and down dale, knapping the chucky stones 
to pieces wi' hammers, like so many road makers run daft. 
They say it is to see how the world was made."
Sir Walter Scott, St. Ronan's Well 1824 


To unsubscribe: send mail to majordomo at wireless.org.au
with "unsubscribe melbwireless" in the body of the message