Industry Forums → Wireless Service Providers > [Equipment] Mesh that works ?

[Equipment] Mesh that works ?

Anyone have any experience with Real and realistically priced Mesh gear?

Back in 2006, I helped a group of students setup a relatively large mesh network in a dense urban and slum type environment. VoIP was the main application, throughput for computers was limited to 256kbps because of limited backhaul.

Price $250 per node: X86 Geode CPU board ; 2X miniPCI radios ; mix of 7dBi omni and 6dBi or 12dBi patch.

For firmware, the students used a variation of “Design and Evaluation of a new MAC Protocol for Long-Distance 802.11 Mesh Networks”, MOBICOM, Aug/Sep 2005.

Ideas:

»commotionwireless.net/docs/supported-devices

UBNT devices are shown, among others.

Maybe bito can pop in and comment on the following:

»wiki.deliberant.com/index.php/Li···n_Manual

»www.wilibox.com/css/hw/WMN-720.pdf

Maybe bito can pop in and comment...
--
Many believe, but few will question...I decline to be blind.

Another option:

»www.engeniustech.com.sg/files/pd···1110.pdf

There was a whole bunch of mesh products that used Intersil Prism based 802.11b PCMCIA cards which were once widely deployed but gradually disappeared.

I know of quite a few smaller mesh networks (5-20 nodes) using RoamAD, Willibox and various other firmware running on Gateworks Avilla boards with upto 4 CM9 miniPCI cards. We also discussed "upcoming" ADI Engineering/RoamAD mesh products back in 2007. »Re: ADI Engineering Mesh Radio Products

SimplySmart ran a small mesh network in Dunrobin (near Ottawa) using Nortel gear and BelAir mesh gear was deployed in Ottawa and Toronto. However those units were $1000 and up.

Mesh that works

"Mesh...that works" are mutually exclusive words. Hahahahahaha

OK, I'll go back in my corner.

Re: [Equipment] Mesh that works ?

"realistically priced", hmmm, not new but you can get them at a very nice priced used!
We have two POPs with Proxim AP4000M, they go two nodes deep. We have about 15 units in two different POPs and they have been running trouble free for years. We have the ruggedized/outdoor units in service but just found these on Ebay »www.ebay.com/itm/Lot-of-5-PROXIM···3705c11c. Warning!, these units will (or have already) benphased out by Proxim.
You will have to put them in an enclosure and get the proper pigtails, they are POE capable. MESH can be used so long as you plan your network carefully. MESH is a tool, as every tool, it has its place-- and time-- to be effective.

I have three business ops. where Mesh is probably the best answer. I don't need user mobility across the mesh but I do need a mesh that would be smart about routing traffic vs. load and latency.

SimplySurf... and that Nortel gear was crap - I know, I was in on testing at the Crystal palace... complete garbage on a shoe string. I felt bad for whats-his-name when he bought into that mess.

Re: Mesh that works

Re: [Equipment] Mesh that works ?

2013 Battle of the Mesh competition ... no survivor mesh protocols from 2003.
»battlemesh.org/Endorsements2013

OLSR is the oldest survivor: OLSR is highly scalable. It runs on community wireless mesh networks with 2000 nodes (Athens wireless network), ~ 600 nodes (berlin FreiFunk.net), Leipzig Freifunk net, ~ 400 nodes (FunkFeuer.at).

Future probably belongs to IEEE 802.11s + 802.11u running on commodity 802.11n routers.

How many nodes do you need? And what is the expected type of traffic?

Our mesh products went EOL some time ago. To summarize, there's a large chasm between what current mesh network gear can do and what customers think they want for a "mesh network" wrt cost, number of nodes, and performance.

This is my thought... lots of hype around mesh but deliverable product is not in line with expectations. I think I'll look at a micro PtP setup maybe.

From the UBNT forum, I posted...

Mesh works very well if you have compelling reasons and full understating how to implement it. Problem is that 99% of the people that want mesh have neither the compelling reasons or understanding of the limitations of wireless technology. Mesh equipment providers have pretty pictures of nodes talking to other nodes across town - the reality is that each node is not likely going to see more than two other nodes.

Historically speaking, many cities have ghosts of failed mesh deployments because it was specified by salesmen to the prospective WISP accounting people, and the engineering people were never brought to the table.