Brought to you as a public service of the Open Spectrum Foundation (Stichting Open Spectrum), Amsterdam - Prague

NEWS

Highlights from Ofcom's license-exempt research

Since the adoption of the Communications Act 2003, the UK's regulator, Ofcom, has had a statutory obligation to exempt from licensing radio equipment whose use is unlikely to cause undue interference. No legal limit is set on the amount of spectrum used by such equipment, yet the 2005 Spectrum Framework Review (SFR) asserted that:

"little additional spectrum will be needed in the foreseeable future [for license exempt use], growing to 7 per cent of the total spectrum... As a maximum, then, an additional 250MHz or so of spectrum might be needed for license-exempt use... Expanding beyond [this amount] would result in unused spectrum..."

Despite (or perhaps because of) such low expectations, Ofcom made the enhancement of opportunities for license exempt (LE) radio activities a focus of their research in 2006. A comprehensive suite of studies was funded, and the final report in the series was released just a few weeks ago. Answers to questions like these are sure to help policymakers around the world:

• How much demand is there likely to be for LE spectrum in the next 10-15 years?
• How much do LE applications contribute to the economy?
• How much more could they contribute if power allowances were raised - or would the benefits of higher power be eclipsed by the cost of dealing with more interference?
• Are the economic benefits greater from application-specific bands or from bands open to any application satisfying a few general parameters?
• Is there some frequency "ceiling" above which LE can be the default policy?

Power to the people?

A consortium led by Scientific Generics Ltd. produced the first report in this series. "The primary objective," according to Ofcom, was "to assess the options for revising the regulation of license exempt applications (including power levels) to promote use particularly where spectrum is currently under used. The secondary objective... is to explore the use of higher powers in licence exempt applications in order to provide wireless broadband in rural areas where other broadband technologies might be unavailable."

However, the scope of work performed by the consortium was actually much narrower, focussing on broadband WLANs based on existing standards in the 2.4 and 5 GHz bands. Other LE applications and bands are only mentioned in passing. But what the report lacked in breadth, it made up in depth. The benefits of higher power limits for WLANs were found to be "quite considerable", and not just in rural areas. Net benefits to the UK of raising WLAN-allowed power in the 2.4 GHz band from 100mW to 10W EIRP were estimated as £443 million during the period 2005-2010, while increasing the limit from 2W to 25W EIRP in the 5.8 GHz band adds £238 million to the economy. (These figures include the costs of dealing with additional interference.) Other power increases were also modelled, and in general, more power yields more benefit. To minimise interference, the consultants recommend linking power increases to the use of directive antennas - the narrower the beam, the higher the permitted power. Consequently, interference costs grow much more slowly than benefits.

Ofcom proposed rule changes based on these recommendations in July 2006. To protect Ministry of Defense operations the power limit proposed for the 5.8 GHz band was 4W (consistent with ECC Recommendation (06)04), while at 2.4 GHz several options were suggested with frequency or geographic restrictions. Public comments supported the 5.8 GHz increase, so it was implemented. But the 2.4 GHz options "gained only limited support even from the communities that might be expected to embrace them," as Ofcom put it, so they were shelved. The UK thus retains its 100mW power allowance long after other countries have opted for 1W or more - increasing the cost and complexity of the WiFi "hot-zones" spreading about the country.

A ceiling on licensing?

The most recent LE research report concerns license exemption as the default policy above a certain frequency. This idea was first suggested in 2002 by the FCC's Spectrum Policy Task Force:

"While it is difficult to say what regulatory approach should be used for millimeter wave spectrum [above 30 GHz], the physics of this band are so different than lower bands as to bring into question most of the fundamental precepts of radio regulation. This results both from the high propagation losses due to gas absorption of radio signals and the ease of building antennas with very narrow beams. While licensing is the general presumption at lower frequencies, the physics of these frequencies appear to justify a de novo approach to considering regulatory schemes on a case-by-case basis. It may well be reasonable to question whether unlicensed use should be a major type of use in these higher bands, rather than one restricted to a small set of bands..." (Final Report of the FCC's Unlicensed Devices and Experimental Licenses Working Group)

The team led by Quotient Associates seemed to reach a similar conclusion - before rejecting the idea of a license "ceiling":

"...approximately 40% [of the spectrum between 30 and 105 GHz] could operate under a licence exempt regime with today's technology, 50% would be suited to a lightly licenced regime, and the remaining 10% would be licenced. However, with further development of licence exempt technology, we expect that a number of these lightly licensed uses could be migrated to licence exempt regimes suggesting that a very significant proportion of future uses could eventually be licence exempt...

"[In addition] all spectrum above 100 GHz could be made licence exempt with little risk of congestion... [but] allocating all spectrum above a selected minimum frequency [to LE uses] would not be the optimum strategy..." (Higher frequency bands for licence exempt applications)

The problem, as they see it, is that above 100 GHz large swathes of spectrum are reserved for passive services like radioastronomy, which need protection even from low-power emissions.

The consultants were also persuaded by a new approach to evaluating spectrum alternatives. "Real options analysis" - a technique inspired by similar calculations widely used in the financial derivatives market - apparently shows that deferring a decision is sometimes the best choice. Quotient's team are among the first to apply this methodology to radio. If Ofcom is convinced by the LE example, this approach could affect other allocation decisions - like refarming the Digital Dividend:

"At the present time there are no practical mechanisms that would allow the market to determine the amount and timing of the release of spectrum for licence exempt uses. And the same is largely true of lightly licensed allocations... Real options analysis can provide a framework for decision making in such circumstances. In particular it demonstrates that:

• "Where there are large uncertainties in the benefits to be had or to be forgone, it can be economically advantageous to wait. Non-use of a band can be the highest value use at a particular time.
• "When demand is uncertain and decisions are irreversible, the optimal allocation at a point in time is likely to be smaller than a conventional net present value calculation would suggest. Hence, smaller more frequent releases of spectrum may be more appropriate than a single large release."

LE economics

Ofcom's use of economic criteria to determine whether an application should be licensed or exempt makes one finding of the Indepen/Aegis/Ovum research into The economic value of licence exempt spectrum especially significant:

"certain LE applications, such as short range radars, RFIDs in retail and public access WiFi, could generate economic benefits for the UK which are substantially greater per MHz of use than the highest value licensed applications, i.e. mobile telecommunications and broadcasting. This finding is an important one for Ofcom to consider in determining possible future designations of licence exempt spectrum...

"[U]se of RFIDs in the retail sector is the highest value application per MHz of bandwidth... [E]ven if the RFID spectrum allocation were increased by a factor of 10 to deal with spectrum scarcity, it is likely that the value per MHz generated by use of RFIDs in the retail sector will still exceed that of mobile telephony and broadcasting..."

One wonders what the proponents of "propertyising" all spectrum will say about this:

Application	Annual value* (millions of GBP)	Value per MHz (millions of GBP)
L I C E N S E D
Mobile telephony	21,786	50
Broadcasting	12,269	27
Fixed links	3883	29
U N L I C E N S E D
RFID in retail	2,478	620
Telemetry	600	300
WiFi public access; Home data networking; Wireless building automation	5,270 395 96	These applications share the 2.4 GHz band: 69
Home alarms	143	29
*Annual value of LICENSED services is based on Ofcom's December 2006 study on the economic impact of spectrum. The annual value of UNLICENSED services is Indepen/Aegis/Ovum's "medium demand" forecast for 2020, calculated in 2006 prices.

[Spectrum Policy - 24 April 2007]