Software defined radio and cognitive radios are changing the way regulators should tackle spectrum allocation, allowing shared use of frequency and even peer-to-peer self-forming networks.
While the technology is not yet here, regulators need to license for the future to prevent a situation where valuable spectrum is locked out and allocated for sub-optimal use. Chief of this is the reclaimed 700 MHz spectrum originally used for analogue TV broadcast which many countries are reallocating today.
Speaking at the Mobile 2.0 expert forum meeting hosted by LIRNEasia and the Pakistan Telecommunication Authority in Islamabad, Helani Galpaya, Chief Operating Officer of LIRNEasia, delivered a presentation on behalf of Dr Payal Malik, senior research fellow at the University of Delhi,who was unable to get a Pakistan visa for the event.
Spectrum is viewed historically as a public resource that is owned by sovereign nations and allocated to parties for various periods of time. It is a public resource that creates very private rights.
Once someone owns spectrum, it can create perverse behaviours. Some might just hold on to unused amounts so that other people cannot have access and strategically prevent entry.
The question is how to allocate it best and use it, generate use and generate government revenue.
The trick is to release spectrum at the right time and to release the right amount of spectrum at the right time. This is not as easy as it sounds. It might have been 1800 at one point, 900 at another. It is not a static thing. Governments need to be aware of what is going on to release appropriate spectrum at the appropriate time.
Large chunks of spectrum are owned by broadcast channels. Telecommunications authorities might not have the authority to take it back and institutional arrangements need to be put in place to negotiate if that frequency should be used for broadcasting, wireless broadband or telecommunications.
In most developing countries, governments are holding vast amounts of spectrum. India has large swathes of spectrum owned by defence and it is unclear if it is being used.
Spectrum allocation can be done in three ways. An administrative approach is where the government decides, and this does not have to be free. The market-based approach uses auctions. Then there is the commons-based approach, which is driven by new technology that enables spectrum-sharing.
An administrative approach can be a good thing, especially to stimulate market entry by giving localised spectrum for free or for very low charges in unserved, rural areas.
However, in less-than perfectly governed countries there is often the problem of non-transparent administrative allocation given for no reasons that technocrats could justify _ friends and family allocation.
"If you have that kind of government, administrative allocation is not the best kind of allocation," she said.
Market-based allocation is what is happening in most countries through auctions. This allows people with money and people with technology to determine where to put their money and capital is allocated much more efficiently.
The problem is that it is often an all or nothing approach. If a country gives out spectrum one time in 20 years, then there is a tendency to over-bid. Secondary trading should be allowed so economically better outcomes are achieved. (Thailand's draft frequency allocation act does not allow secondary trading of frequency).
A commons-based approach means that frequency is not exclusively owned and many players can use the same frequency. There are various models of commons use, such as the way different Wi-Fi connections can coexist in a room, having agreed on standards and power output.
The real benefit of commons will be with new technology that can do not just point-to-base, but also peer-to-peer. A device can carry others' traffic, so the more devices added to the network, the more efficient the network becomes.
There is a notion that commons needs to be free, as Wi-Fi is free, but that is not true. It is possible for an operator to bid for frequency and then sell access to a commons created by the owner of a band. This model can still generate revenue for governments.
Dr Malik's research centred on India and how its central allocation model has failed. India gives out spectrum based on number of subscribers, not actual use. This has led to a perverse incentive where operators simply add more users to gain more spectrum and denies spectrum for more economically value data usage. India has no policy for allocating more than 10 MHz chunks (needed for next generation dual-channel 3G) and has run out of more GSM 900 spectrum to allocate.
India decided in September 2006 to auction off 3G frequency but is only starting to move forward today. However, what remains unresolved is access to 2G frequency for any new bidder. Because the bid is only for 3G frequency, it means that it is highly unlikely that any new players will enter the market.
The United States freed up the 700 MHz band which used to be used for analogue television which is one of the best spectrums for wireless broadband. The UK is doing the same. Google said it would bid $4.6 billion for the spectrum if the FCC agreed to four conditions: Open applications (no blocking), open devices (no locking), open services, and open networks based on reasonable and non-discriminatory (RAND) terms.
In this region, it is taken for granted that the Sim is decoupled from the phone. But in the US and much of Europe, most phones are subsidised and under 18 to 24-month contracts.
This is important because mobile operators can provide preferred content if they control the pipe and the terminal device. They can go into partnership with a certain provider and give preference, in much the same way Apple controls the iPhone ecosystem.
The bargaining power is unequal. Operators can choke bandwidth for applications they do not like, such as Skype.
However, that did not happen and the incumbents, Verizon, AT&T and T-Mobile got the block. They got $9 billion (291 billion baht) but did not change the rules or the market dynamics.
The other story from the United States is the use of cognitive radio, or software-defined radio. With frequency management in the software, it allows for devices to sniff the air, see which frequencies are available and adjust itself to start using a free band.
Cognitive radio would allow for the reuse of white space. In the US there used to be huge guard bands between analogue broadcast channels. After the switch to digital, guard bands are still needed but with much smaller sizes. Cognitive radio can use these unused bands where they can.
Google is trying to set up a national database. When a cognitive device turns on, it will consult with this database, find out which frequencies are available and then register itself with the database.
"Are we going to wait another 10 years? Think of 1985 and Wi-Fi. We had this attitude _ what are you talking about? Unlicensed spectrum shared by multiple users in the room. Today we take Wi-Fi for granted and it's very cheap.
"The moment India adopts a sensible policy for 700, or China gets into developing extremely cheap cognitive devices, this vision will become a reality. With this in mind, regulators need a cohesive view of spectrum today," she said.
Galpaya said it would be too late if regulators went with old-school allocation now on 700 MHz as it could not take spectrum back for at least another 10 years. It would be better to regulate for the future and prepare for a commons-based approach.
She also called for a move away from administrative models, to try for market-driven models and to decouple licenses from start-up spectrum.
Finally, Ms Galpaya said that there needs to be better harmonisation with ITU bands and for countries to have long-term, 15-year, spectrum plans to help in building road maps.