First I need to correct an assertion I made last week in a blog.  I always hate having to admit to a mistake, but the one thing I hate more is someone not owning up to a mistake when the evidence is clear.  I spend a lot of time analyzing (and debunking if necessary) the work of others and it frustrates me to no end if they ignore the evidence so I would be a hypocrite if I behaved in the same manner.

I wrote a blog last week criticizing Spectrum Bridge and the city of Claudville for wasting white space mobile spectrum on a backhaul network, but it is becoming clear that I made a mistake based on what Peter Stanford from Spectrum Bridge has told me.  It turns out that Spectrum Bridge was not using their White Space spectrum for backhaul and I had misunderstood the statements from their Chief Marketing Officer Rick Rotondo, so I must apologize for my mistake and clarify what is really happening there.  However, I do stand by the other parts of the article opposing unlicensed spectrum usage of White Spaces and I will reinforce some of those points below.

The key point that I was trying to make in the original post on Claudville was that mobile spectrum should not be wasted on backhaul networks, but Spectrum Bridge in Claudville was actually using their White Space spectrum on a last-mile access network.  It also turns out that they are using the 200 MHz range (174-216 MHz specifically) White Space spectrum (much further from the 700 MHz upper boundary than I had thought) which is more appropriate for connecting homes and businesses with fixed radios than for mobile applications.  That’s because at 174 MHz, the antennas would have to be 16″ for a quarter-wavelength and that’s just simply too large to fit into the lid of most laptops.  400 MHz and above with 7″ or shorter antennas would be more realistic for mobile applications.

Specifics on the Claudville White Space broadband network

I had a nice chat on the phone with Peter Stanford and he explained that the Claudville implementation uses an experimental license from the FCC (which grants them temporary exclusivity for roughly the next two years until the part 15 unlicensed rules for White Spaces are likely to be finalized).  The radios are housed in a weather proof enclosure outside the home or business with an external antenna and the fixed base stations are limited to 4 watt EIRP.  Claudville is situated near some major roads that have fiber running through them so Spectrum Bridge taps the fiber along these roads and then runs a White Space access network for half a mile to two miles to the home or business.  This 200 MHz radio link provides the last-mile broadband connection but the user can employ wired or wireless Ethernet technologies like 2.4 GHz 802.11 Wi-Fi for the in-home or in-business connection.

The radios themselves use an 802.11g radio chipset that is very similar to the Wi-Fi devices we have running in the 2.4 GHz range and they are configured to use a single 6-MHz unused TV channel getting anywhere from 2 to 6 Mbps (depending on signal strength and quality) of total effective bandwidth per base station.  I asked Stanford what happens when White Spaces becomes Part 15 unlicensed and anyone can buy a cheap White Spaces access point and they start abusing the spectrum by running maximum power over the same frequencies.  Stanford acknowledged that these are potential problems but said that these rural areas have roughly 200 MHz of unused TV White Space spectrum so it’s less likely there could be a problem.

Still, I raised the issue that some people could potentially just grab a lot of that spectrum while using maximum power levels just because they can.  Someone might just use White Spaces as a quick-and-dirty backhaul network just to avoid having to build a 10-foot tall tower for a line-of-sight backhaul and that could easily cause problems for this unlicensed broadband network.  Someone else might just decide to use all of the available White Space spectrum to video monitor their farm animals and they don’t care if their signal propagates out 2 miles degrading someone else’s network.

An Ofcom report showed that unlicensed Wi-Fi has gotten so bad and over saturated in London that 90% of the spectrum is occupied by beacon and other housekeeping overhead which means the unlicensed 2400 MHz network is only 10% usable for actual data.  With 200 MHz spectrum which propagates much better than existing 2400 MHz Wi-Fi, it is conceivable that consumers will start buying “super White Space Wi-Fi” devices in the 54-698 MHz range that are going to a mile or more in range.  That would likely mean a much higher occurance of the kind of severe problems seen in London with unlicensed 2400 MHz Wi-Fi.  This will be something that Claudville will have to worry about in the future when White Spaces becomes unlicensed and it is one of the main reasons I have always been wary of turning White Space into unlicensed spectrum.

Suburbs and urban areas will have much less unused White Space spectrum available and there will be more users wanting to use the scarcer spectrum so it could be even more problematic.  Then there is also the potential of miconfiguration and broadcast TV interference.  Stanford explained that there is also the potential of using White Space networks with licensed 54-698 MHz band from the secondary spectrum market so it is possible to lease spectrum for licensed broadband applications.  That also allows for much higher powered radios well beyond 4 watt EIRP which means it would be possible to cover huge regions with lower population densities.

Lastly, Stanford explained that the White Spaces network is also used by the utilities for Supervisory Control and Data Acquisition (SCADA) and that these narrowband applications can be allocated 25 KHz or 50 KHz of spectrum so that they’re not wasting several MHz of bandwidth.  As an Information Security professional, the prospect of utility companies using wireless networks for supervisory control of power infrastructure scares me but that’s a discussion for another day.