Original URL: http://www.theregister.co.uk/2011/04/06/specnet/
Microsoft shows how to crowd-source spectrum management
You do have a spare analyser, don't you?
Researchers from Microsoft have proposed a new way to monitor spectrum usage, by connecting up idle analysers and providing an API allowing anyone to make use of them.
The concept of SpecNet, as the researchers term their creation, is to link up spectrum analysers around the world and allow them to be accessed through a standard API so that a company, government, or individual, could run a query to see who is using what and where.
But the complexity is in putting that all together; creating software to manage programmed queries and keeping it simple enough for general use, while maximising the application of spectrum analysers, which are expensive bits of kit rarely used to their full potential.
SpecNet addresses the technical problems by creating APIs of different complexity dependent on the class of user – so the high-level API could be used to establish the occupancy of a band within a specific geographic area, while a different API is used to monitor for transmissions of a specific shape or power. SpecNet takes care of distributing the task to the right analysers and gathering historical data, while dealing with the fact that analysers will disappear for extended time periods while the owner is using them.
The researchers present answers to these problems (in PDF/557 KB form – really interesting but quite specialised), as well as addressing the balance between granularity, fidelity and time that spectrum-scanning demands (the more carefully you look, the more you'll see). What they don't address is the political side of the problem.
A worldwide network of spectrum analysers is a nice idea, but as the researchers admit, it's not easy to convince labs and research establishments to connect up their $10,000 bits of kit to the internet for everyone else to use. Even if that can be addressed, there's the problem that many analysers spend their lives in basements or Faraday cages (or even down salt mines) specifically to avoid picking up local radio signals. But ignoring the political practicalities for a moment, it is worth thinking about how useful SpecNet could be.
A national map of radio usage in the UK would cost about £2.11m annually to maintain, according to CRFS, which carried out a trial run for Ofcom in 2009. That proved too rich for our regulator's blood, but did attract the attention of the FCC, which plans to spend more than $10m to create a map of the USA as part of the country's National Broadband plan (mainly to aid the search for more spectrum to sell off).
The Microsoft chaps do hopefully suggest "governments may be willing to sponsor a set of spectrum analyzers dedicated for SpecNet use", and it would certainly seem to be an effective alternative to CRFS's method of strapping spectrum analysers to the roofracks of travelling salesmen.
As we exploit radio spectrum with greater efficiency, there is more interest in knowing just how efficiently we are exploiting it, and if there are any bits we've missed; SpecNet might not be the answer, but it's an interesting step towards it. ®