Nine companies are bidding to run the US white space database, but only Google plans to give access away for free - which should ensure the race doesn't last long.
The FCC plans for exploiting white space - unused TV spectrum - require the creation of a database showing which frequencies are available and where. Nine companies have submitted plans for running that database, along with technical details and a business model to fund it; Google's submission simply says the search giant will shoulder the cost itself and give away the service for free, which is going to make life hard for those charging for access.
The original idea was that "white-space devices" would detect existing users and switch frequencies to avoid them, but that's technically dubious and hasn't been demonstrated to the FCC's satisfaction. So the proposed solution requires devices to locate themselves then connect to a database which will allocate a frequency along with a timeout, after which the device will have to repeat its request.
Leaving aside the problem of how a wireless device connects to the database before it has a frequency allocated (perhaps over a slower connection such as cellphone), the FCC has said that devices will have to connect to an FCC-approved database before they'll be licensed for sale. To that end submissions were invited, and nine companies responded before the 4 January deadline.
The proposals are largely occupied with proving their ability to run massive databases, work with radio spectrum and/or work with the FCC. No one has ever run a radio system like this, so no one can really claim experience in the area (though most of the proposals try). The FCC foresees mobile devices, which might require new frequencies every few minutes and number in millions, all of which will need to be tracked and provisioned in real time - which should prove interesting.
The proposals are split between those who favour a centralised data store, and those who think peer-to-peer is a better way to synchronise data. Equally divisive is the question of device security, with some proposals favouring proper certificates for every device, while most think a shared-secret approach would be more practical. But these are mere details compared to the question of how the service is to be funded.
Only one proposal, from KB Enterprises and LS Telecom, suggests charging end users an annual fee, and even then it's just one option listed. Everyone else, excepting Google, believes the best way is to charge manufacturers of white space devices a one-off fee for an indefinite licence to use the database. The FCC is expected to approve four or five databases, so they can then compete on price and value-added services.
At least, they could if Google's proposal didn't say the search giant would happily fund the project from its own pockets for the five years specified by the FCC:
"Google has sufficient funds and access to capital to develop and operate the proposed database for a full five-year term... Google has no current plans to rely on user fees."
We contacted the other companies to see how they planned to respond to a competitor giving away their product. KB Enterprises were the first to respond, explaining that unless something changed "it will be impossible to compete with Google for this business".
Others were less apocalyptic, but equally concerned. "Google’s presence here does indeed complicate the business model," Comsearch told us, adding: "We think there will be a market for database providers as an alternative to Google."
But why would the kind of far-eastern manufacturer expected to develop white space devices choose a value-added supplier when consumer electronics are so price sensitive? Fear of Google, according to hosting-hopefuls Spectrum Bridge: "Some, if not most of the companies that are taking prominent roles in the white spaces discussions... may not want to partner [with Google] for obvious reasons."
So Microsoft, Dell, Motorola and their mates will force manufacturers to avoid Google 'cos they don't like Google - not something we'd want to bet our business plan on, but Spectrum Bridge has other ideas too. "They may also want to collect non-confidential info [and] may not want Google providing or collecting this data on their behalf."
The database host will know where users are and the kit they're using, both of which are commercially valuable pieces of information. Google thinks that data will pay for the database, and Google is very good at extracting value from information; but even if it can't turn white space into gold, it will have five years to drive the competition out of business.
The proposals were all filed before 4 January, and we're now in the 30-day period during which anyone is free to add their own comments. Following that there'll be 14 days for the proposers to reply before the FCC closes the doors to consider who's going to get the deal.
There's no time limit on that, but the FCC will probably anoint four or five of the proposers as "officially approved white space database suppliers" with whom device manufacturers can then register and start selling devices - though how many of those manufacturers will choose to pay for the privilege remains to be seen. ®