Original URL: http://www.theregister.co.uk/2008/01/05/new_spectrum_licences/

Why there will never be another GSM

Ofcom squares up to EU over spectrum carve-up

By Bill Ray

Posted in Mobile, 5th January 2008 07:02 GMT

Comment Ofcom is changing the way spectrum is licensed in the UK to remove usage restrictions, but the EU is calling for region-wide harmonisation of use to create economies of scale.

These diverse approaches are going to lead to an unavoidable clash of ideologies that could mean the success of GSM is never repeated.

Great chunks of spectrum are becoming available around the world with the switch off of analogue TV, and traditional licensing mechanisms are increasingly being seen as limiting innovation while protecting incumbents and preventing proper competition. But usage-specific licences are an easy way of minimising interference and ensuring equipment compatibility.

When broadcast radio started in the UK in 1922, it took five months for the 18 companies involved to come up with a proposal for a national broadcaster to manage transmissions, and to decide that both receivers and transmitters would need to be licensed.

Back then, the choice of frequency was just a matter of picking one which offered the greatest range and was able to carry an audio signal. Any potential interference was addressed by switching off transmissions every seven minutes and listening for complaints coming in on the same frequency.

Crystal clear

A general rule is that the higher the frequency the greater the data that can be carried, but the shorter the range and the greater the power needed for transmission. So audio broadcasting was quick to grab Medium Wave, around 530 kilohertz to 1.6MHz, allowing the BBC to reach half the UK population with only eight transmitters at limited quality.

TV needs a lot more bandwidth, but still wanted decent range so broadcast TV grabbed the next available chunk, from 474MHz up to 850MHz, leaving later applications such as mobile telephony to camp in higher frequencies starting around 900MHz.

When spectrum was simply being allocated, without charge, governments generally grabbed all they could for their own use. In the UK that amounts to over half of the spectrum below 15GHz, mostly (75 per cent) used by the military, while 12 per cent goes to civil aviation and five per cent to the emergency services. The rest is split between maritime and scientific uses.

Not only is Ofcom planning to auction off great chunks of the spectrum currently used by analogue TV, but it's also set its sights on that 50 per cent owned by the various government agencies, including the military, arguing that such agencies should pay market prices for spectrum or be forced to sell it off to ensure greater utilisation.

Those who buy spectrum in the upcoming auctions won't have to hand them back after a limited time period. These will be perpetual licences which can be sold on or sub-let to other users. And those users won't be limited to a specific technology. They'll be able to deploy whatever they like as long as they don't generate too much interference for their neighbours.

The success of GSM can be attributed to the way a particular technology was mandated, along with the frequencies at which it operated; initially 900MHz, allowing mass production of equipment which drove down the cost of handsets as well as network infrastructure. The use of the same technical standard on neighbouring frequencies (licensed by different operators) also reduces the chance of interference, as the standard incorporates limits on broadcast power as well as interference-avoidance techniques, making life much easier for the regulator.

The problem with this approach is that it forces the regulator to guess which technology is going to be effective over the period of the licence, and licensees to gamble that the regulator is right. With GSM the regulator hit the nail on the head but, in the UK, T-Mobile, Orange, 3, and O2 all paid for 5MHz blocks of spectrum (from 1.9GHz to 1.92GHz) to deploy Time Division Duplexed 3G services, but no such services were ever developed and even now the spectrum stands empty.

An even more extreme example, arguably, is the EU's proposal for a frequency to be mandated for DVB-H services - broadcast mobile TV. Not only is DVB-H a controversial choice of technology, but the market for broadcast TV to mobiles is very unproven and the EU could end up with a chunk of spectrum in which it would only be legal to operate an unpopular service using a redundant technology.

Commission for interference

The alternative, and that proposed by Ofcom, is a Spectrum Usage Rights (SUR) allocation which will specify how much interference a frequency licensee is allowed to generate in neighbouring frequencies, and how much interference they can expect to experience in their licensed band.

Interference comes in two types: in-band and out-of-band, the former being when a neighbour signals leak into your band, the latter being when the strength of their signal overwhelms your receivers (such as when you drive past a Southampton radio transmitter and it becomes impossible to receive anything but Radio Solent until it's out of sight). Winners of next year's UK spectrum auctions can expect to receive SURs which peg that interference at specific levels, but allow the licensee to do anything with the frequency.

These changes aren't limited to the few well-publicised national licences; there are over 40,000 licences that Ofcom manages for fixed wireless connections - often used as backhaul for mobile phone networks, but also to connect companies together. These too will move to being tradable, and non-technology specific.

This is in stark contrast to EU proposals for mandating technologies to specific frequencies in an attempt to emulate the success of GSM. The EU is committed to a more flexible approach, but while Ofcom considers itself to be leading the trend, the EU commission sees it vanishing over the horizon and is looking for powers to bring it back in to line.

Everyone agrees that spectrum licensing needs to be more flexible, but opinion is divided on how flexible and how quickly. The EU is riding a tide of good publicity on the back of its popular cap on international roaming charges, and while Ofcom tried to show what a consumer champion it was by fining companies over the premium-rate scandals it's not secured the public trust as yet.

Memories of the UK 3G auctions still loom large over any spectrum allocation, and this new way of allocating spectrum is going to be hard to explain to most punters who will just see it as more money going into the treasury.

The EU's new-found love of PR will no doubt be mobilised to sell harmonisation, though Ofcom's determination combined with the UK public's distrust of anything coming from Brussels is likely to see that campaign ultimately fail and Ofcom's free-market approach dominate the future of spectrum licensing.

Anyone really interested in how Ofcom views the future could do worse than read a copy of Essentials of Modern Spectrum Management. This rather dull book claims to be a guide to how spectrum is licensed around the world, but reads much more like a manifesto of how the authors feel spectrum ought to be licensed. And given the authors' close relationship with Ofcom, those views can easily be seen in Ofcom policy. ®