Carriers vs cops: Australia's spectrum conundrum
New services or emergency services?
Australia, like other countries
subsidising the broadcast and consumer electronics industries rolling out digital TV, is now preparing to auction the old analogue TV spectrum for the best price possible.
The obvious answer is a spectrum auction, the obvious customer the telco industry. That has the tech press wildly imagining a new golden era of cheap broadband, somehow believing that the same carriers that price a mobile gigabyte at up to $80 will become more generous with the LTE services they're going to launch over their new, expensive spectrum.
What should happen to the "digital dividend" (which everybody except Treasury seems to regard as some kind of social description rather than a purely economic one) is a lively debate, and currently centres on whether emergency services should have any of the newly-available spectrum.
The emergency services argument, mostly prosecuted by the police (but presumably cheered on quietly by fire brigades and ambulance services), is that there will be plenty of spectrum to go around. The switch to digital TV will make 126 MHz of spectrum available; emergency services just wants 20 MHz of this.
The counter-argument, prosecuted by the telecommunications industry and mostly supported by the tech press, is that the whole 126 MHz has to go on the auction block to maximise the number of new services Australia will get.
The carriers' case is being ably prosecuted by John Stanton, head of the industry's representative body, the Communications Alliance. His position is that if the industry gets the whole of the available spectrum at auction, emergency services organisations can then negotiate with the industry for access to the spectrum.
But who is right?
The telcos' position: spectrum equals LTE
Today's mobile networks portion the spectrum out in 20 MHz blocks. While technologies like LTE-Advanced will work in a 20 MHz channel, higher performance implementations – the maximum downlink speeds above 100 Mbps that get the tech press so excited – need wider channels, like 40 MHz.
Hence the clash with emergency services. To allocate 40 MHz each to three bidders needs 120 MHz: keeping back 20 MHz for emergency services leaves only room for two 40 MHz blocks (with 26 MHz as the odd-man-out leftover).
If they can persuade the government to preserve the entire digital dividend for the auction, mobile telcos Telstra, Optus and VHA (the merged Vodafone-Hutchison entity) are the winners: there's really nobody around to bid against them.
It's unlikely that an emerging fourth mobile carrier would find investors to pitch a brand-new network against the majors. Investors would need to back a company not just to win the spectrum auction, but to roll out the whole vast infrastructure of a new, competitive mobile network – and to survive the cash burn of acquiring customers in a relatively mature market.
In other words, the industry position in this debate is really the position of the three mobile carriers that dominate the market.
Emergency services' position: we need more than voice
The position of the emergency services is simply that their existing radio networks support almost nothing but voice communications.
Perusing the Australian Communications and Media Authority's radio licence database supports this view. NSW Police, for example, has mostly 10 kHz and 16 kHz chunks of spectrum in the 400 MHz band. Whether it's important or not, it's certainly accurate for them to say that this spectrum doesn’t support any kind of high-speed data.
That spectrum constrains the applications and the content they're able to use: computer apps have to be able to live with very low data rates, and video just can't happen.
Without new spectrum, emergency services have only two options: abandon any ambition to deploy high-bandwidth applications, or gain access to spectrum held by carriers.
Can't we share?
The industry answer – one that's supported by much of the tech press in Australia – is the second: as proposed by John Stanton, emergency services should simply negotiate spectrum access with the carriers, perhaps under some kind of regulatory regime.
There are, however, problems with this proposal – more, perhaps, than are immediately obvious.
First, let's look at the modes of communication emergency services use. While it's true that all three arms – police, fire, and ambulance services – are heavy users of mobile phones, these only provide one-to-one communications. Emergency services also need broadcasts over (moderately) open channels.
While mobile technologies support broadcast modes, emergency services would vastly prefer that their broadcasts weren't over a public network. I would suppose that carriers would also prefer not to have to worry about whether they could support the broadcast requirements of Australia’s various state police, fire and ambulance agencies, along with federal agencies.
There's also the question of access arrangements: under what circumstances would the emergency services be able to require that carrier networks are turned over to them to cope with an emergency?
At one end of severity, I can easily imagine that using an LTE device to transmit video to a coordination centre is sufficient for police and ambulance to handle the fallout of a bad road accident. However, what level network access would suffice for (say) a major cyclone, bushfire or flood?
At some point, emergency services would argue – probably reasonably – that their operations take precedence over private communications, and that they therefore require exclusive network access.
Exclusivity has its problems, though. Drawing the line between minor and major emergencies is just one. There's also the question of what communications individual citizens need in an emergency.
The combination of personal communications and social media, while perhaps less important than the massing of emergency resources, has proven powerful in emergencies, both to help disseminate real information about conditions on the ground, and to help people cope with emergencies (for instance, by checking on friends and family without needing to wait for central authorities to find and report on missing persons).
If an entire network were turned over to emergency services, the citizens' interpersonal communications is lost; yet that seems a more-than-likely scenario if spectrum is only available to the telcos.
Finally, there’s a much simpler question of the infrastructure itself.
The media was astonished at how well mobile infrastructure can survive some kinds of emergencies – the recent Brisbane floods and North Queensland cyclone, for example – but resilience is relative.
From an emergency services point of view, a mobile base station looks very interdependent and fragile: it can't survive long without mains power, and it can't communicate at all without backhaul.
If a disaster knocks out some base stations, others nearby can take up the load – but at the cost of bandwidth (since speed falls with increasing distance) and congestion (since more users are crowded onto the same tower).
The emergency services' preference for managing their own networks is both prudent and comprehensible. Their need would likely be for fewer base stations than a mobile network, each covering a greater area. Even sub-megabit network speeds would look lavish by comparison to today's voice-only networks, and they know how many personnel any given base station needs to support.
So if I ignore what looks, from my vantage point, to be media barracking for the telecoms industry because that's the point of view we most readily recognise, the debate appears almost irreconcilable. Emergency services can't compete with telcos in an open auction, and even if those carrier networks could serve all their purposes, citizens would probably object to emergency services' requirements, even if carriers didn't.
If it weren't for the fact that I don't know enough about the constraints of spectrum allocation, I would argue that the problem is administrative. Why was Australia's TV spectrum redrawn so that the digital dividend has an odd-man-out at the tail of the spectrum?
This, however, ignores the complexities of spectrum policy. The ACMA has to reconcile engineering, budget and political demands of various powerful lobbies (broadcasters as well as carriers).
Which leads to a question: did the government make too many concessions to TV broadcasters in the digital TV changeover? In the era of "new media", people are watching less TV than they did formerly, yet part of the digital TV tradeoff has been to give the broadcasters extra channels.
Those extra channels increasingly seem like dead ducks. Instead of exciting new content, they seem to be turning into rerun channels or the home of programmes that don't suit the "flagship" channels.
Perhaps the emergency services' needs would be better served by a rethink of the needs of broadcasters rather than a stoush with carriers. ®