The agenda for spectrum
Martin Cave and William Webb, authors of a new book on spectrum management, set out trends in usage and an agenda to improve spectrum use
The challenge for the regulator remains the same as always – to maximise the value of the use of radio spectrum through ensuring that the highest-value users have access to the
spectrum and that the interference between them remains controlled to an optimal level. However, the solution to this challenge is constantly changing as usage patterns change and new tools to manage users develop, often based on changing technology. Here we look at the key trends relating to spectrum usage and suggest where we think spectrum management will evolve in the coming decades.
Trends in usage
There are hundreds of different uses of the radio spectrum and trends and changes in many of these. However, where the users require relatively small amounts of spectrum, or spectrum for which there is little competition (eg. satellite use of frequencies above 20 GHz) then these have limited impact on spectrum management approaches. We discuss what we believe are some of the most important trends relating to those uses that generate the highest economic value (broadly cellular and broadcasting), those that compete for the most congested spectrum and those where innovation might be most likely to occur. The trends are not presented in any particular order.
Trend 1: Growing importance of unlicensed spectrum We are using WiFi and Bluetooth ever more. Home WiFi networks now routinely have tens of devices connected to them whereas a decade ago they might have had one or two. Monitoring devices are slowly appearing around the home, such as WiFi-connected security devices, and many more are being shown at trade shows. Other systems, such as the home energy system, may make use of a separate network within the home to connect gas and electricity meters, in-home displays, and eventually smart appliances. Out of the home we are increasingly looking to log onto WiFi hotspots wherever we find ourselves. Bluetooth accessories are proliferating, with many health- and fitness-oriented devices being developed for wearing and devices like Google Glass being tested. As this usage grows so will the possibilities of interference, while at the same time our tolerance for not being connected will fall. The use of 5 GHz spectrum will offer some relief but regulators may need to become more inclined towards making bands unlicensed than they are at present.
Trend 2: Ubiquitous connectivity will become ever more important Once connectivity is widespread then behaviours change. People make greater use of cloud servers and plan ahead less on the assumption that they can retrieve the information they need as and when it is needed. Less material is stored locally, enabling devices without hard disks. All of this makes ubiquitous connectivity ever more important, fueling a circle of change. This is perhaps less about rural coverage and more about ensuring that all of the devices that a person is carrying can connect wherever they happen to be.
Trend 3: Ever higher data rates will become less important Each new generation of mobile system (2G, 3G, 4G) has been predicated on approximately a tenfold increase in data rates compared with the previous generation. This has led to requirements for ever broader bandwidths and ever more mobile spectrum. But we are now reaching a point wheremany homes and users have more than adequate data rates for all that they could wish to achieve. It is hard to envisage why we might want yet another tenfold increase. This is causing a rethink as to whether fibre to the home is a necessary solution, and it will bring into question
It may be that the seemingly insatiable desire for spectrum for mobile usage will come to an end.
whether data rates moving towards 1 Gbps are necessary for 5G. This in turn will change the demand for spectrum for future mobile data systems. Coupled with the trend toward ever more use of WiFi it may be that the seemingly insatiable desire for spectrum for mobile usage will finally come to an end.
Trend 4: The internet of things (IoT) will grow rapidly IoT is predicted to grow from perhaps 1 billion connected devices in 2015 to over 50 billion by 2025. This rapid growth will require some new spectrum, which regulators have not yet identified, and will provide a new source of value from the radio spectrum.
Trend 5: The categorisation of usage into fixed, mobile, and broadcasting will become of less relevance At present regulators divide usage into fixed, mobile or broadcasting. But these are increasingly converging. Broadcast and mobile systems will tend to share platforms and spectrum in the future and broadcast-only solutions will increasingly be switched off.
Trends in spectrum management tools
A range of new tools will become available for spectrum management. These will include the following trends.
Trend 6: Dynamic spectrum sharing will become commonplace As white space deployments grow and sharing moves into other bands, the concept of shared opportunistic access will become more widely accepted and it will be considered normal for virtually all spectrum bands to be shared in some manner.
Trend 7: Regulators will stop predicting interference and start monitoring it The complexity of predicting what interference might occur will continue to grow while new tools will become available to rapidly fix interference where it does occur. As a result, regulators will increasingly deploy systems up to a point at which interference is observed and then set this as the licence conditions. New ways of monitoring interference will be needed to allow for real-time reporting across a range of users and uses.
Trend 8: Regulators will seek more ways to encourage innovation Regulators often aspire to stimulate innovation but rarely achieve it. Innovation is becoming ever more important in government agendas and regulators will start looking at mechanisms to provide favoured allocations for new approaches and other novel ways to encourage innovation.
Trend 9: Receiver performance will become an ever-growing issue and regulators will eventually move to control receiver performance as well as transmitter performance The performance of receivers often prevents or limits new technologies or uses in neighbouring bands. Regulators will increasingly realise that some specification of expected receiver performance is needed in order to effectively manage spectrum.
Our agenda to improve spectrum use
Throughout our book we discuss areas where current approaches to management and use of spectrum fall short. In such cases we have put forward suggestions as to how they might be improved. The five most important proposals of this kind are set out below.
1 Move to a position where (almost) all licences are shared The coexistence of increasing demand for spectrum and evidence that many valuable frequencies are underused has been one of the drivers, over the past five years, for more spectrum sharing. This process has been assisted by the development of new real-time technologies for dynamic spectrum sharing which allow multiple users to coexist. These methods have augmented earlier temporal and geographic sharing of bands and the limited use of unlicensed spectrum.
It is time for these possibilities to be reflected more fully in rights of access to spectrum by the replacement of exclusive licences by arrangements which allow access to multiple users, possibly on a hierarchical basis which gives some users priority over others. The result to be expected is much greater flexibility in use of spectrum and lower prices of access to it. This could be accomplished by a process of progressively replacing exclusive licences with less restrictive alternatives, introduced in ways which manage the associated risks. One effect might be the emergence of intermediaries which offer flexible access to multiple clients. We recommend in the future a brisk increase in the number of licences recast in this way, even if in practice some of these will continue to be exclusive.
There is mounting evidence that inefficient use of public spectrum is increasing relative to commercial spectrum. We have outlined a strategy for dealing with this but its success hangs in the balance in many countries. Public/commercial sharing of spectrum may in some countries be a more palatable means of switching spectrum to more efficient uses.
2 Link licences to interference generated rather than power transmitted Current licences are not suitable for a world where usage can be changed and bands repurposed. There have been many examples of this from the interference caused by Nextel in US public safety systems in the 1990s through to the more recent cases of LightSquared in the US and TV interference concerns in white space. The basic problem is that a licence that specifies the amount of power a transmitter is allowed to emit does not fully constrain the interference caused since this is also determined by the density of transmitters and other factors such as their height.
A better approach is to specify in the licence the interference that can be caused and then allow the licence holders to configure their transmitters in such a way that this interference level is not exceeded. Studies such as Ofcom's 'spectrum usage rights' have pioneered more suitable licence formats that achieve these objectives while being practical to implement. However existing players have generally resisted their introduction due to increased complexity and the problem that the benefits most immediately fall on the newer entrants rather than established companies. Regulators should look at the wider benefits for usage of radio spectrum rather than the needs of incumbents and adopt licences specified in interference terms as the default across all spectrum bands.
3 Manage receiver performance There are two aspects to the performance of a radio receiver – its ability to decode the wanted signal and its ability to reject unwanted or interfering signals. While the former is typically a core part of radio design the latter is more problematic. Manufacturers may not know what interfering signals will be experienced either now or in the future and have strong incentives to design the lowest cost receiver by taking a low expectation of interference.
If such poorly performing receivers are widely deployed, for example in millions of TV sets, this can cause the regulator to prevent the introduction of nearby spectrum usage that might impact these receivers, despite the fact that they are substandard. Knowing that this outcome is likely allows receiver manufacturers to design to ever lower standards resulting in a 'too big to fail' problem.
To some degree this problem is linked to licence conditions (see above) in that a clearer specification of the interference that spectrum neighbours could cause would clarify the receiver requirements. There are many issues here associated with national licensing versus global manufacturing, the role of standards bodies and more. It is clear that this is a serious problem that is progressively causing greater delays in
optimal use of spectrum and needs addressing. Regulators need to work together on a global basis perhaps through the ITU or similar bodies to ensure that receivers have appropriate specifications and that these are implemented.
There is greater scope for regional collaboration if frameworks can be found that do
4 Move away from previous ITU labels to a new categorisation Bands are currently divided into categories such as 'fixed', 'mobile', and 'broadcast' at international level and have been for nearly a century. However, the boundaries between these uses have been increasingly blurring. For example, many now view broadcast content on mobile devices, downloaded via cellular or WiFi. There are mechanisms such as 'broadcast modes' within cellular networks which allow broadcasting to be performed from those cellular networks. The labels were originally intended to facilitate interference management by placing the highest-power networks (broadcasting) into some bands and lower-power networks (mobile) into others.
But as discussed above this method of managing spectrum is suboptimal. From a management perspective it is the interference levels generated, rather than the service deployed, that is most relevant. The categorisation also tends to limit innovation by persuading users that they cannot, for example, change use to a mobile application in a band that has been categorised as broadcast, whereas in practice this particular example would likely work acceptably. A better division would be 'low', 'medium' and 'high' interference levels, ideally linked to specific limits and new licence types. The idea of reclassification should be discussed internationally by spectrum managers and others, options considered, and a new approach adopted at ITU level and cascaded down to regional and national plans.
5 Reconsider regional and global spectrum management There is greater scope for regional collaboration if appropriate frameworks can be found that do not limit innovation. Regional bodies should study their role and look for where they can add additional value. This is an issue of fundamental importance which deserves careful and disinterested study at a very high level.
Spectrum management has come on a long journey through its 110 year history. Technological developments have led to rapidly growing demand and also to innovative options for using spectrum based around sharing and reactive assignment. These developments have exposed the weaknesses of the legacy command and control approach to spectrum management. This led to the development of market-based management in the past 25 years, initially focused on auctions but later extended to secondary markets in spectrum and also to spectrum pricing.
With the notable exception of auctions, the market-based methods have not been widely applied, and have not yet delivered all of the expected gains. The world where regulators step back and allow the market to function still looks to be in the distant future. Making the best use of spectrum appears to need a mix of new technology and new regulation. The most prominent area of application of this combination presently resides in spectrum sharing, based on techniques such as dynamic spectrum access.
Other developments which have strong potential are the implementation in licences of new ways to control interference, and the extension of regulation into domains such as receiver performance.
As spectrum becomes an ever more critical component of our lives the difficulty of making such changes grows. The risk remains that cautious regulators forever remain one step behind technology and need. Our book has set out how we believe they and others can best act to put the airwaves to work for the benefit of the economy and society.
The prize is enormous.
Martin Cave is a regulatory economist with extensive telecoms and spectrum experience. He is a visiting professor at Imperial College Business School, London, and a past member of Ofcom's spectrum advisory board. William Webb has led the design of Weightless, a new standard for M2M communications. This article is adapted from their book, Spectrum Management, published by Cambridge University Press. See bit.ly/1jxxq7q
spectrum for the good of society
Policymakers and regulators now have a choice of two excellent new books about spectrum. In the last issue of Intermedia (January 2016), we reviewed
Understanding Spectrum Liberalisation, which covers the extent to which liberalisation policies have succeeded over 'command and control' models. The authors of that book say the
free market model is currently a disappointment, in not being applied widely to broadcasting, not providing enough mobile spectrum and not stimulating much trading outside the US. But while there are major constraints, there are promising options such as in spectrum sharing.
Martin Cave and William Webb cover much of the same ground in Spectrum Management – the technologies, ways of allocating spectrum, pros and cons of trading and sharing, international and political issues, and much more. But their book takes a deeper dive into economics and policy, as would be expected from Cave, who is a regulatory economist and telecoms specialist. The book has the subtitle 'Using the airwaves for maximum social and economic benefit', and is a necessary update of the author's previous Essentials of Modern Spectrum
Management, which seems like a lifetime away now back in 2007.
Throughout, the new book has an eye on the complex balancing acts needed in spectrum management, and sets out clearly what the functions of national regulators are, from determining
and changing the use of spectrum, to assigning and policing it, and conducting forward looking studies to enable allocation to adapt. They say that spectrum currently contributes about 3% to a country's GDP with 'third best practices' – and while it is not practical to get perfect efficiency, we can at least move to a 'superior second-best outcome'.
There are two primers, on technology and economic challenges, with the latter setting up the bulk of the book by noting that allocating spectrum is extremely complicated, given a large number of bands, many known and unknown issues, and demand for uses growing at variable and unpredictable rates. It is no longer possible to think
of solving allocation centrally and the authors set out the modes of access – rules-based such as public and private commons, and licence-based, and a combination of the two.
They cover ways of allocating spectrum – auctions of course, but also liberalising how licensees can use their spectrum, spectrum trading, and what they term administrative pricing, which is where the regulator sets prices and can use tools such as opportunity-cost and 'business model' pricing, the latter
'mimicking' the calculations that a company would do. The pros and cons of allocation methods are set out.
Detailed chapters on auctions and spectrum trading follow, trading being one way of sustaining market forces after
a primary allocation, but with modest success so far, and possible remedies to competition concerns are discussed.
Spectrum pricing and valuation also gets a chapter, before the book enters a third part, on emerging spectrum management approaches such as sharing, which is the 'defining feature of recent work' and the authors say there is a clear need to increase flexibility, as set out in their agenda article.
Chapters also include many examples and break-out boxes, and there are also plenty of references.
There's plenty more – on interference, the 'struggle' for the UHF band, public sector use, TV white space and so on – and a chapter on the wider economy, presenting various estimates of spectrum's economic consequences – but they say more research on mobile broadband impacts is needed.