1. Home
  2. Wireless Networking
  3. Locustworld Mesh Networking

Locustworld Mesh Networking

Just started reading about mesh networks for wireless 802.11b access. This Locustworld stuff is pretty amazing/interesting.

Open source/ free software - linux based <surprised ?>

Load it on most any PC with 64 Meg memory and one of a selected list of wireless 802.11b adapters (only works with those having linux drivers) and you get both a) a wireless access point serving 802.11b clients and b) a mesh network.

The mesh network is a network of these boxes. They dynamically seek other mesh boxes out and connect. Network changes are recognized dynamically. Any client attached to the mesh can get to the internet through any internet connection that is connected to any mesh node. And a mesh node can funnel internet requests to its directly connected internet connection, or pass such requests on to other nodes in the mesh.

Software free. (did I say that already)

Preconfigured boxes about $400 or so.

formatting link
formatting link
Weatherresistent boxes designed for outdoor use $430 + antenna.

In basic models, one MeshAP can service the mesh and serve as a wireless AP simultaneously on a single 802.11b card. Add a second card or add an ethernet connection to a dedicated AP such as D-link or Linksys, and one radio serves wireless clients while another one serves mesh traffic. Might be useful for a mesh with many clients and few internet access connections.

Monitoring and modification all via the internet

Apparently you can upgrade the software via the internet also.

Look at

formatting link
for more info. In the tradition of Linux, it is somewhat difficult to find answers to your questions and good synopses. But I can suggest the following:

formatting link
snipped-for-privacy@lists.bawug.org/msg01451.html
formatting link
formatting link
formatting link
formatting link

Main LocustWorld Americas site:

formatting link

I am now looking at my PCs and those of my family to find out which has a wireless card that is supported. I downloaded the software and it seems to boot just fine (bootable cdrom) but no wi-fi adapter working.

If you are using this I would like to hear from you. If you know of other great sources of info, please post them also.

Reply to
Bob Alston
Loading thread data ...

Single radio mesh networks suck because the same packets get sent over the air at least twice on the same frequency. Install a maze of these and you'll have many times the same packet being resent. At least dual radio systems have a chance. I don't wanna get into routeing algorithms, but methinks you'll find that mesh networks require some rather sophisticated routeing.

MIT Roofnet:

formatting link
Champaign-Urbana Community Wireless Network:
formatting link

Reply to
Jeff Liebermann

Here is an interesting link to a NE Texas implementation that appears to be of about 20 nodes

formatting link

Reply to
Bob Alston

One more. Meshbox:

formatting link
formatting link
I know of several other similar mesh networks. They all have the same general characteristics. They work just fine when they're small and have less than about 100 nodes in a single "airspace". They start to fall apart at about 20 nodes if they use single radios in each poletop. Those with two radios per poletop do well up to about 50 nodes. If they have an intelligent routeing algorithm, they scale to about 200-300 nodes. Rooftop Networks (Nokia) did most of the pioneering work. They went through two hardware mutations and gave up. Metricom/Ricochet is a another early mesh, but differs in that the client radios do not act as part of the mesh. I have some experience with Ricochet and note that managing such a large network was non-trivial. One does not need to be efficient when one is small. "Self-Healing" and "Self-Configuring" were only buzzwords at the time. I'm not sure if there has been any real progress in these areas. Anyway, Metricom went from single band 900MHz ISM poletops, to 3 band poletops (900MHz ISM, WDS band for backhaul, and 2.4GHz ISM) as soon as they discovered that store-n-forward repeating on a single channel just didn't scale.

The mesh idea is really attractive for municipal networks as it eliminates much of the cost of an expensive backhaul.

formatting link
course it pollutes the users bandwidth with excessive packets, but that's not really a consideration as it only becomes a problem with large systems. Sell the system, take the money, it works, and run. Then, when it grows and becomes constipated, it's someone elses problem (or replace everything with the latest model).

Since mesh is sufficiently popular to get the attention of the IEEE, the 802.11s committee is working on standards. If they can't solve some of the inherent problems with mesh networks, at least they can identify them and reduce the hype.

In my never humble opinion, only poletops that have multiple radios and a decent routeing algorithm have a chance of scaleing. See:

formatting link
formatting link
are others but these are the only two I can remember.

Have fun with your experiments. Mesh can be made to work. But, if you're serious, always keep in mind whether what you're doing will scale into a larger system without hitting some limit, inefficiency, or complication. For example, running out of routeable IP addresses or can't get IANA to hand out some more?

formatting link

Reply to
Jeff Liebermann

A "never humble" opinion is as good as any opinion. <grin>

Reply to
Bob Alston

Cabling-Design.com Forums website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.