On Tuesday Brett Glass gave a lunch talk entitled Lessons from Laramie: Broadband Innovation on the Wireless Frontier. The following is a summary of the talk based on notes that I took. The full video is available here and his slides are available here if you want more information after reading my summary.

Disclaimer:

This write up is based on notes I took during the talk. It is entirely possible that I got some of the details wrong. I’ve probably missed some important points as well.  A video of the talk has now been posted to the Berkman web site . I’m posting my notes in hope that they will be useful for those who don’t have the time to watch the full talk and as a way to help people determine if the talk is something they would be interested in watching. This summary is not intended to be a substitute for watching the full talk.  Also, I’m a computer scientist but not an electrical engineer nor an expert on networking. It is entirely possible that I’ve misunderstood important aspects of radio transmission, the electro-magnetic spectrum, or other technical issues.

Brett Glass talked about his experience building and running LARIAT an ISP in Laramie, WY. LARIAT is notable because it was arguably the world’s first WISP (terrestrial, Wireless high speed Internet Service Provider). LARIAT was founded in 1992 as a 501(c)(12) non-profit co-op to serve unserved/underserved areas in and around Laramie, WY in 1992. Glass took LARIAT private in 2003 at the request of the membership. In terms of geographic coverage area, LARIAT has been growing by about the size of the island of Manhattan each year.

The big advantage of wireless Internet for a rural area such as Laramie is the low deployment cost (around $100 per square mile). LARIAT uses residential roof mounted antennas to provide Internet access. They offer houses in prime locations such as hill tops free Internet in exchange for allowing LARIAT to mount roof top antennas which are used to provide Internet access to other homes.

Glass discussed the legal and regulatory environment LARIAT operated in and bemoaned policy decisions made in the absence of science. In particular he argued that policy needs to be made based on Shannon’s Law. Shannon’s Law says that capacity increases linearly with bandwidth but only logarithmically with signal to noise ratio. Essentially this means that while very crowded spectrum is difficult to use efficiently, having exclusive access to spectrum is only marginally better than lightly shared spectrum. Glass’s problem with the current spectrum allocation regime is that there’s really only unlicensed spectrum and exclusively licensed spectrum. Exclusively licensed spectrum is currently prohibitively expensive to obtain and unlicensed spectrum is polluted by consumer devices. He argued that we need a non-exclusively (“lightly”) licensed spectrum which is not polluted by consumer devices and unburdened by needless regulatory requirements. Shannon’s Law tells us that lightly licensed spectrum would be almost as good as exclusively licensed spectrum and vastly better than unlicensed spectrum that is crowded by consumer devices.

Glass detailed the reasons why a small ISP such as LARIAT is essentially forced to use unlicensed spectrum. Currently the FCC only auctions off spectrum rights for very large geographic areas. For example, there were auctions for the spectrum rights to all of Wyoming but not for the right to just Laramie. Another problem according to Glass, is that the foreclosure value of spectrum always exceeds it’s utility value. In other words, the value of the monopoly rents from blocking your competitors is likely to be greater than the ability to profit from the spectrum directly. This is a flaw in our current auction process.

The problem with using unlicensed spectrum such as the 2.4 GHz band to provide Internet access is interference from consumer devices. For example, Glass described an incident he called the “Biki Problem”. A customer reported problems with Internet service only on bright sunny days. The problem turned out to be the customer’s daughter sun bathing with a 2.4 GHz cordless phone in front of the antenna for wireless Internet.

Another issue that LARIAT faced was high bandwidth cost. The Incumbent telecoms charge LARIAT 10 times their cost for bandwidth but LARIAT has no other option. Backbone providers would charge a minimum of $15,000 which would mean that in a town of 28,000 LARIAT would need near 100% market share.

Glass made a number of policy recommendations. With regard to spectrum, in addition to increasing the geographic granularity of spectrum licenses and creating non-exclusively (lightly) licensed spectrum, he bemoaned the fact that much of the spectrum in Laramie is currently unused by it’s licensees. But the licensees are unwilling to rent out this unlicensed spectrum to others. He suggested applying the doctrine of “adverse possession” to spectrum or allowing “homesteading” on spectrum. Essentially if owners of spectrum are not willing to use it productively “squatters” should be free to make use of it. He spoke out against unnecessary regulation and net neutrality regulations in particular but did advocate fixing the broken “middle mile” (special access) market for upstream bandwidth.

Glass’s objection to network neutrality was interesting particularly since most of the Berkman community supports network neutrality. He was concerned that one of the plans that LARIAT offered would not be allowed under come of the current net neutrality proposal. This plan offers customers a guaranteed 256 kbs connection for $30 but does not allow them to run a server and certain file sharing programs. (256 kbs may sound slow but keep in mind that this is a guaranteed minimum. Glass said that the actual speeds will bounce up to 512 during times light times but will never drop below 256 even during “prime time”. By contrast Cable and DSL providers usually advertise a maximum speed which you would be unlikely to obtain in practice.) Glass pointed out that LARIAT was able to offer this plan because it could buy asymmetrical bandwidth cheaper from its upstream providers and that the restrictions on file sharing programs were necessary ensure that bandwidth usage was asymmetrical (e.g. download traffic was larger than upload traffic).

In the Q&A, I suggested that WISPs were a rarity and asked Glass if his policy recommendations – particularly with regard to net neutrality – made sense for areas such as Cambridge where there was only essentially a single choice for Internet. He replied that WISPs are more common than most people realize. Many WISP’s hide to avoid being squashed by big ISPs and a lot of WISPs don’t need to advertise. Glass cited D.C. Access — a WISP that operates in Capital Hill – as evidence that WISPs can and do exist in urban areas. Unfortunately, as far as I was able to determine Comcast seems to be the only option for those of us that live in Cambridge.

I found Glass’s description of LARIAT interesting and his policy recommendations on spectrum seem to have merit. However, I’m unconvinced by his arguments against net neutrality. The sad reality is that for most of the country the market for high speed Internet is a duopoly at best and a monopoly at worst (as in Cambridge). Until we have a truly competitive Internet the potential danger from traffic interference is simply too great not to have net neutrality protections.

UPDATE: 23 Feb 2010.  I contacted CWISP on Brett Glass’s suggestion in the comments to see if they could provide home wireless service to me in Cambridge. I received a reply asking for my address and the type of service I was looking for.  I provided that information but received no response.  I send a follow up email after a week for still received no response.