Tips for creating a community wireless mesh?

First off: I know Jeff doesn't particularly favour wireless meshes (for good reason), nor do many others. I'm going to assume that you're going to establish one anyway.

Are you certain one doesn't already exist? Most cities have a mesh project of their own already up and running - check to see if there is one, and if so inquire about joining.

Basically, on the hardware side, you're going to need your own "base" then provide services from this (ie. Internet, streaming radio/video). And then you're going to need to get nodes to transmit these services over a wide area. The way BrisMESH did this in one instance was to get an educational institution whose premises are on a large hill behind the project. In the end they let the mesh put antennas up on the top of their roof (near the commercial antennas) and those are capable of reliably transmitting up to 1km away in each direction, given LOS.

Hardware will be quite expensive in the end, so if you're doing this on your own for the community's benefit it doesn't look like it's worth it. You'll be needing a server, an AP (WRT54G's are favoured because you can put custom firmware on them easily) or two, and then antennas. You'll probably need to mount the AP's outside too, on your roof perhaps (if this is in a residential situation).

Then your nodes are going to be needing at the very least an AP and a good antenna. You're also going to be needing good, strategic places to put your nodes. Use Google Earth to find possible sites - you'll also need to analyze the elevation, LOS and Fresnel clearance.

That leads on to the next issue, community involvement. Have a look at the BrisMESH node database -

actual Brisbane mesh (and most of the nodes on it) were established well before my interest in the project, but a database such as that one can really help you work out potential nodes and their availability, as well as help "recruit" folks to your mesh. So you let people add themselves to the database (they become interested in letting you put their equipment on their property) and then you work out how to get them connected.

I hope I've given you a bit of an idea of what you need to do or at least pointed you in the right direction for some examples of what others have done.

All the best.

By all means hire a professional for this.

EVEN the years of experience, it is not a trivial deployment for even a professional.

For starters:

Free or advertisement biased subscription won't work, as Earthlink and others have figured out.

• Figure an AP on ever third lightpole if you aren't using outside mounted antennas an ,000 for each light pole mounted AP.

• If outdoor mounted equipment, figure on 0 per subscriber your cost, minimum.

• Figure on 0 per T1 that will support only 100 casual users.

• Figure on a full time IT professional for support.

• Figure ,000 for a CALEA solution if you have paying subscribers if your not comfortable rolling your own. And a mandatory 24x7 "SPOC" phone (single point of contact).

Check out

Adair

They rather upset a few of their customers earlier this year.

That sucks, had not noticed that.

Adair

From fiercebroadbandwireless.com edition Thu, 20 Dec 2007

[quote] Meanwhile, the muni-WiFi market, which hit the upper curve of the hype cycle in 2006, came crashing to reality as providers realized there isn't much of a business case in offering free access to residents. The technology's limitations became apparent. WiFi isn't meant to be a high-scale network, and the cost to blanket a city is more than what was ever anticipated.

1. The muni-WiFi market implodes

Headlines across the country announced the implosion of the muni-WiFi market as EarthLink, one of the service's largest cheerleaders, pulled back on its initiatives and some cities scrapped their WiFi plans. EarthLink is now looking for strategic alternatives for the business, aka sale, and several municipalities have canceled planned projects. The problems: Usage is low in many cases, scalability has been a challenge, devices and capabilities have been found to be wanting and business models promising free community access have proven not to be be viable, for the most part.

Is this a sign that the muni WiFi market has failed or is the industry settling into a more rational period of deployment based on securing cities as anchor tenants rather than free WiFi access for the masses? We are seeing examples of muni-WiFi done right when the service is used to better municipal services.

Daniel Pitts hath wroth:

Wireless mesh sucks.

Go unto:

and read. It's a bit dense, but I think you can get the idea that speeds are low, thruput sucks, capacity is limited, and error rates are sky high. Mesh networks don't scale but that's not covered in the test conclusions. See:

See:

for a mailing list that covers wireless ISP's. Yes, you're building an ISP (internet service provider), with all the political, social, economic, and business complications of running an ISP, with the added enjoyment of a totally unreliable means of delivery.

I can offer more specific denunciations of mesh after you get some numbers together. How big a community? What level of service are you expecting to deliver? Who is paying the bills?

Incidentally, the "social aspect" of a WISP is primarily answering the support phone, at odd hours, from irate customers. Running a WISP is a great way to convert friends into irate customers.

You could add this to your list as it also covers dual and multiple radios.

Sure. Ummm... what list?

Thanks for the white paper link. I hadn't seen that one. BelAir was one of the first vendors to sell dual radios, that don't have the performance problems associated with single radio repeaters. I just skimmed the PDF but it looks like they have the benefits fairly well covered. However, they presented a large number of graphs without the slightest clue as to how they were derived or where they were borrowed from. I can't even tell if they're test results or MathCad/Matlab simulations.

I just found the 2004 Philadelphia muni wireless mesh network business plan at:

The techy gibberish is at the bottom. I've read it several times and fail to understand what they're doing (or trying to present). Anyway, it's worth skimming if you're considering a large system.

More on Philadelphia:

I gotta admit they're off to a fair start with coverage:

15 square miles currently covered. Only 120 square miles to go. Oh well.

I was feeling lazy and couldn't be bothered to start a new thread for the OP for a one line entry. I presume he is making a list of things to read.

I am in the process of clearing out a few hundred bookmarks and I had BelAir listed in several locations. BelAir was

I think they have deliberately kept their papers as simple as possible. Building Municipal mesh:-

Minneapolis seem quite pleased with theirs.

David Fairbrother hath wroth:

Yep. Nobody listens to me.

Mesh has a few compelling benefits, that make it look attractive. It's relatively cheap and easy to deploy as the backhaul costs are a big piece of the total cost. If rapid routing reconfiguration is designed into the mesh software, it's theoretically possible for the network to repair itself in case of localized failure, jamming, vandalism, or power outage. If hardware and operations costs are of paramount importance, it's possible to ignore the disadvantages of mesh, and concentrate on exploiting the benefits.

A bit under 2000 users but only 11 users (and a total of 21 nodes) currently connected:

Fairly nicely documented. I wouldn't mind seeing some performance and efficiency statistics, but that's mostly for the geeks and nerds. The status page is rather crude:

but seems useful.

Kinda looks like your Wiki has been vandalized:

Don't click on anything at the top of the page as it tried to download an ActiveX control. Yep, just today:

Sigh.

I guess nobody saw the "First Citiwide Change Bank" skit on SNL, then?

"We just make change."

-- Les Cargill