No im not falling for that trap. Your example used a 100% node to node
connection rate, that is not reasonable for wireless.
2 to 5 radios/node is. This reduces the network update messages. My idea
with a 3 radios is 1 BH AP, 1 BH CPE, 1 Client
AP. The BH CPE should be smart enough to know where nodes are physically
and be able to have the BH CPE jump from
one of the 3 to 5 nodes it can see(this is optimal in my opinion). Cell
sizes should be small with short high capacity bh's. This
is just a temp solution till the wisp can string copper to connect the
APs. Pure wisps simply will not survive with out wired
infrastructure beyond the upstream. To this end mesh like Lonnie has
included is a good step when used correctly. Locust
World mesh is ahead of the meshing game but there Wianna totally
detracts from it.
Jeromie
Matt Liotta wrote:
My example used wireless P2P links, which has no inherent weakness
over fiber P2P links from a topology point-of-view. It would appear
you are falling into the same trap as others by forcing mesh to be
something it is not. Mesh is just a network topology; no more, no
less. Sure it is possible to come up with specific examples of
wireless-based mesh networks being terrible ideas, but that doesn't
mean there is anything wrong with mesh itself. I would argue that in
almost all cases the topology is not what is at fault.
-Matt
Jeromie Reeves wrote:
There is a very big difference from fiber mesh and wireless mesh.
Wireless is classicly a bunch of HDX links
where fiber is PtP links. Your example doesnt make it clear that the
difference is what cause's 802.11[a|b|g]
mesh "suck" and fiber/copper mesh's "not suck". The solution is multi
radio units that can select peers based
on more then just essid (channel, hop count to the edge, packet loss,
ect)
Jeromie
Matt Liotta wrote:
Attached is a quick rundown of basic mesh theory that I put together
in light of the recent thread. It hasn't been peer reviewed or
edited, which I would normally do before sharing publicly. But since
I only wrote because of a thread on this list I figured I would just
share it. Feel free to pick it apart.
I do want to point out a couple of things though. First, this was
written in a generic way only covering mesh as a theory. As written
it can be applied to various transport technologies from fiber to
wireless; though I do provide an example using wireless P2P links.
Applying mesh theory to wireless P2MP or ad-hoc networks would
require special coverage.
-Matt
--
WISPA Wireless List: wireless@wispa.org
Subscribe/Unsubscribe:
http://lists.wispa.org/mailman/listinfo/wireless
Archives: http://lists.wispa.org/pipermail/wireless/
