re: Nortel: Another Fine MeshMixing 802.11a meshed backhaul with 802.11b access is a nice trick but how can you hope to deploy such a network on a commercial basis in an urban area using the normal WLAN unlicenced bands?
Imagine, some poor operations team setting up a backhaul network using say 200m spans, they test it and everything works. They announce coverage to the commercial guys and the first paying (remember this gear is to be used by players like ATT and BT!) users come on line.
10 days later someone switches on a nice big corporate WLAN in between two of the backhaul APs and bang - the entire zone (not just one AP!) goes off line.
Yes a mesh might offer alternative paths but that just means that either a) the backhaul capacity drops by 50% with the first corporate WLAN in the zone or b) everything is fine for the first one but you loose the lot when the second system comes online near the alternative route.
Compare this to a DSL or licenced spectrum WiMAX backhaul with 802.11a/b/g access - the big corporate WLAN nearby will only kill (part of?) one AP's coverage not the entire city.
Boring network engineering problem I know but I'd hate to see he business plan with all the "re-visit, re-engineer" OPEX costs...
re: Nortel: Another Fine MeshAgreed. Esp if the 802.11a backhaul in adhoc mode hits say Proxim Tsunami in UNII beam or even one of those OrthagonalModulation jobs from WiLan. Most of these thingz are point2point for sure but yah, its a matter of the interference mitigation process where the MESH Integrator would then have to go knocking on doors to ask to share the band etc etc. Can you Engineer a totally interference free MESH-nope, not in our lifetime but I betcha each ISM band user could mitigate the effects of band crowding for sure. I just started training on Firetide-« product and same issues are being addressed, with some areas _left_blank_. Also, why is Nortel late to every dance? Do they have 7250/72xx ready now, or is still in trial, and they are just being quite careful? Or is this MESH idea of backhaul just so so new and fresh that the stuff really needs the shakeout? Interesting to wait and watch- I'll check back on this forum later. John R. in Canada
re: Nortel: Another Fine MeshIf I understand the stated problem correctly, Collision Avoidance built into 802.11 does not interfere. A better term would be coexist. It steals bandwith, not clobbers packets, or cripples data transmission. In a mesh network, since all nodes are on the same frequency they "interfere" with each other in the same way.
re: Nortel: Another Fine Meshno not really. the big issue is short links always beats long links and that when you engineer a network you normally make the links as long as possible according to the constraints you observe which means NEW short links that pop up (or drive past) will kill these long links
Imagine, some poor operations team setting up a backhaul network using say 200m spans, they test it and everything works. They announce coverage to the commercial guys and the first paying (remember this gear is to be used by players like ATT and BT!) users come on line.
10 days later someone switches on a nice big corporate WLAN in between two of the backhaul APs and bang - the entire zone (not just one AP!) goes off line.
Yes a mesh might offer alternative paths but that just means that either a) the backhaul capacity drops by 50% with the first corporate WLAN in the zone or b) everything is fine for the first one but you loose the lot when the second system comes online near the alternative route.
Compare this to a DSL or licenced spectrum WiMAX backhaul with 802.11a/b/g access - the big corporate WLAN nearby will only kill (part of?) one AP's coverage not the entire city.
Boring network engineering problem I know but I'd hate to see he business plan with all the "re-visit, re-engineer" OPEX costs...