Ofcom unveiled proposals earlier this month to auction off radio spectrum reserved for military and emergency services traffic, in order to meet growing demand for mobile broadband. It has committed to releasing 500MHz of public sector spectrum by 2020, in a bid to deliver faster mobile broadband services to the ever-increasing cohort of UK subscribers that are eating up data on a biblical scale.
In practice the spectrum will be auctioned off in a ‘flash sale’ which is expected to raise between £50 million and £70 million. But is this the best approach and what does it mean for the telecoms marketplace? What if, instead of parcelling up small chunks of spectrum and bolting them onto mobile operators’ existing allocations, which the current plans would result in, the spectrum were released for unlicensed use?
Rather than resulting in marginal new benefits to users, opening the spectrum up to unlicensed use would mean that when a mobile operator needs to multiply capacity in a certain area, it could shift users onto the unlicensed network, enabling it to carry more traffic and provide a better and more consistent service for users. The abundance of spectrum and flexibility to use it would result in enormous capacity gains for everyone.
We’ve already seen RAN sharing and tower sharing agreements that have turned into complete corporate mergers. The predominant focus for mobile operators now is in managing cost of every delivered Megabit. With the huge increase in the volume of data being delivered, and continuing to grow at CAGR over 50%, this focus is understandable.
Tower sharing is one solution. Distributed antenna (DAS) solutions are another, for larger indoor and campus deployments, where each operator hooks up to the DAS with their preferred base station and shares the antenna path with their competitors. These solutions are big plays – the capex and delivery costs are big, the project timelines are big and the cost of change is big, so they only really work in the largest deployments (usually a few hundred thousand square feet or more).
For most typical sized applications, small cells have an alternative. There is a 3GPP standard feature called MOCN (multi-operator core networking) which is an ideal technical solution for allowing operators to share their RAN infrastructure but from within a small cell, not at the antenna or tower. But there’s nothing in the feature itself which says how one operator can make money by allowing its competitors access to its spectrum, nor how they can apply policy to the sharing arrangement.
That is up to the pressure growing as existing spectrum fills up. There is another spectrum model on the horizon which illustrates the value of expanding accessible spectrum: which is to deploy LTE in unlicensed spectrum, or LTE-U. In LTE-U, or at least one version of LTE-U called Licensed Assisted Access (LAA), a radio carrier in licensed spectrum is aggregated with another in unlicensed spectrum. So, when a mobile operator needs to multiply up its capacity in a certain area, it can shift its users on to the unlicensed carrier, and just carry more traffic, to the delight of its subscribers, and its shareholders.
When people talk about LTE-U they’re normally talking about an unlicensed carrier in the 5GHz band – sharing the same frequencies that WiFi 802.11n or 11ac uses. There’s so much spectrum up there that the capacity gains can be enormous, and there’s plenty to go around – up to 700MHz in some areas.
The fact it’s unlicensed means that operators can set the carrier frequency freely within the band to avoid existing WiFi APs and other military radar stuff that still lingers up there, and still have room for multiple streaming HD videos per user. There are some regulatory and standardisation hurdles still to be overcome; and the handsets don’t yet exist to exploit that spectrum in LTE but this could change rapidly and the green shoots are there.
Although it sounds a long way off, it isn’t. In the context of the upcoming Ofcom auction, Band 40 (at 2.3GHz) is already in use in Asia. Its use is already standardised by 3GPP. The handsets exist already to use this spectrum and they’re in wide circulation.
So, what would happen if Ofcom adopted an unlicensed model for the newly released military spectrum? Lots of good things. Existing operators would be able to deploy infrastructure (using any of the RAN sharing techniques described earlier) to give the optimum network performance for their customers. It would become an early test bed for the alternate LTE-U model, where new operators deploy infrastructure, like WiFi APs, but re-using their LTE solutions in the unlicensed band.
All of the great quality of service, small packet performance, high user density, high mobility performance of LTE would be available instantly using phones that are already on the market. By making the whole of the spectrum available as a single 40MHz block, the freedom of configuration to avoid interference would be maximised, and by trunking the whole spectrum into a 40MHz block instead of dividing it up into separately licensable 5MHz blocks, the capacity and quality of the new spectrum would be maximised, to the benefit of all – operators and end-users alike.
So why has Ofcom licensed this spectrum to continue the present model, even though the world needs RAN sharing in small cells? If mobile broadband is the imperative, and it is, then why choose a model that divides the spectrum up into sub-optimal, inflexible chunks? Isn’t it time to set the scene for an entrepreneurial new generation of players who can use the spectrum as well, and see who wins out on a level playing field?