[MLB-WIRELESS] Out with the hobbyists, in with the dancing elephants

[Donovan Baarda]

On Tue, 2003-12-09 at 08:36, Ryan Abbenhuys wrote:
> I don't like it.
> 
> Unless they develop some fantastic new wireless technology that offers a
> huge amount of bandwidth and can spread itself across a large part of the
> spectrum I don't see it working.

hmmm sounds like spread-spectrum :-)

I thought that the limitation of how much spectrum you spread yourself
over was legal, not technical.

> At the moment I just don't believe it's possible to get 50 nodes in 1
> square kilometre linked in a mesh on the same frequency, yet still offer
> high bandwidth.  Each node would just be shouting down the others in an
> effort to be heard and result in nobody being heard.

I'd be wary of saying "not possible"... I've said it a few times about
things I thought I knew pretty darn well, only to be shown up for a fool
by others much smarter than me.

I'm sure if you had a map of which nodes could see which nodes, and
which couldn't, you could build a fancy time-division multiplexing
protocol where nodes could calculate and/or reserve time-slots for
talking to other nodes that would minimise collisions but still allow
other "hidden" nodes to communicate at the same time.

The other thing to remember is despite the distributed nature of the
"mesh ideal", most real world applications have a central server or two
that all the other clients talk to. The reality is each client can only
be guaranteed (server bandwidth / number of clients). The important
thing is to maximise the server bandwidth. This is an assumption that 
could simplify things.

-- 
Donovan Baarda <abo at minkirri.apana.org.au>
http://minkirri.apana.org.au/~abo/


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