Google’s Brilliant Proposal

If Google’s proposal to the FCC for 700MHz auction rules is adopted, the radio spectrum in question will be both more valuable to those who acquire it and less expensive for those who use it.  Think this is an impossible contradiction?  Just think Internet.  Also remember the concept of “oversubscription” and keep in mind that most frequencies are not used most of the time.

Google requests that the FCC “clarify” that those who purchase spectrum can resell by means of either a dynamic realtime auction (something they know a lot about) or by selling off the right to use a tiny slice of frequency over a tiny range as part of the cost of buying the radio used to access the spectrum or a combination of both.  Devices would contend for the frequencies they use just as WiFi radios or users of the wired Internet contend for transmission space.  Prediction: telco engineers will say that this chaos will lead to unacceptably poor quality: same thing they said ten years ago about the Internet itself and a generation before about cellular technology.  Push your disregard button.

From the Google filing:

“As has been pointed out by various studies, the vast majority of viable spectrum in this country simply goes unused, or else is grossly underutilized. Our nation typically uses only about five percent of one of our most precious resources, and even that minimal use is inefficient compared to what is technically possible today…. Furthermore, the airwaves can provide huge economic and social gains if used more efficiently, as seen today with the relatively tiny slices used by mobile phones and Wi-Fi services. Additionally, in some cases, while the legal rights to use the airwaves have been allocated and assigned, networks are not yet built out. This situation constitutes an entirely avoidable waste of valuable spectrum…”

Some of Google’s evidence for underutilization comes from studies done by Shared Spectrum Company . The company makes the kinds of radios needed to share spectrum so certainly has an interest in this.  One impressive study (warning: 20meg plus download) they performed as part of a National Science Foundation project.  It was conducted in a place of heavy airwave usage (New York City) during a time of presumably peak usage (the 2004 Republican convention).  Overall spectrum usage during the test was observed to be 13%.  This number includes broadcast channels which are in use around the clock.

Usage is low because the owners of most spectrum don’t use it most of the time and, except in the slivers of unregulated spectrum used by WiFi and Bluetooth and the like, we don’t share spectrum.

Shared Spectrum Company and others have developed radios which can operate in an environment where each radio has to dynamically share the spectrum with other radios and be prepared to recover from collisions when two or more radios try to broadcast at the same time on the same frequency.  This dynamic sharing of spectrum means much higher use of the available resource than allocating spectrum to specific users for their exclusive use even though they won’t use most of it most of the time.

If a company were to buy some of the spectrum at the FCC auction and then resell it as Google suggests, they would be able to support oversubscription – e.g. they could sell the right to use the same spectrum many times over confident that the statistical chance of everyone trying to use all the spectrum they’re entitled to use all of the time is nil.  This is exactly the way that ISPs sell service today.  It is not at all uncommon to sell two dozen users each two megabits per minute of access even though the actual pipe which connects the whole group of two dozen users to the Internet only has four megabits of capacity (if that).  This isn’t fraud; this is why Internet access can be affordable even though never wholly predicable.

Google says that its proposal would be particularly good for rural areas since there would be relatively few claimants and therefore low prices for this spectrum which is particularly needed in these same rural areas (Vermont, for example) that don’t have enough landline access.  I think they’re right.  But the price of access in urban areas will also be lower than it is today because, in urban areas, use of the spectrum for Internet access will be competing with relatively abundant landline access.

Everybody wins?  Not quite.  Better use of radio spectrum for Internet access is not good news for those who are overpricing and underproviding landline access today.  The Google proposal is no shoo’in at the FCC.