3. Spectrum: Invisible Infrastructure
Across all geographies, one of the biggest digital infrastructure constraints in the coming years will be the availability, allocation and use of mobile spectrum – the bands of radio waves over which data and voice communications (as well as other over-the-air media) travel. This constraint is also one of the most complex.
Spectrum, by definition, is a limited resource. The amount currently released is far less than that required to support the expected growth in mobile data traffic, which increased 80% to 1.5 exabytes a month by the end of 2013, according to Cisco10, and is expected to soar by a factor of 1,000 in the next 10 to 15 years.11 Technology has helped overcome similar constraints in the past, and it will no doubt continue to do so, but governments and operators also need to do more to alleviate issues of availability, allocation and harmonization that constrain the ability of various participants in the mobile ecosystem to invest in infrastructure and deliver services.
Availability, Allocation and Utilization
Governments are releasing and redirecting additional spectrum for mobile use; they need to hasten these efforts. Many are planning to do so. A significant amount of spectrum is not fully utilized: valuable bands in the 600-700 MHz range are currently inefficiently employed by television broadcasters, for example. Many bands reserved for government and military use are not being used all the time. In some cases, operators are not fully using their spectrum holdings.
While they have the clear goals of providing value and delivering spectrum to the entities that will use it most efficiently, spectrum auctions do not always function as intended. The priorities of ministries responsible for finance and ICT can conflict. Some auctions have become exorbitantly expensive both because spectrum is scarce and because some governments, with a short-term focus, covet the cash that they raise. Companies nonetheless feel bound to participate, lest they lose access to resources they need, but in some instances successful spectrum purchasers find that they lack the capital to build out the infrastructure necessary to put the spectrum they have purchased to use. Governments that focus too narrowly on budget goals may also lose out on larger opportunities to stimulate economic growth through the release of licensed and unlicensed spectrum.
Over the long run, approximately 25% of all “capital investments” by network operators are dedicated to acquiring spectrum, a level that can limit funds available for investments in new infrastructure.12 Moreover, high-spectrum costs have an impact on operations. CSPs cannot always buy the bands that would be most efficient given their current holdings. Fragmented holdings lead to greater complexity in operations and increase costs for equipment (both network and handset) manufacturers.
Meeting the spectrum needs of large and small players by imposing restrictions and incentives can be a tricky balancing act for most governments. Large companies tend to need more spectrum owing to their bigger subscriber bases; their experience and customer bases also help them use spectrum more efficiently. Smaller, sometimes disruptive, CSPs can be the source of new business models and other innovations. Reserving spectrum (or too much spectrum) for entrants without experience can reduce overall availability and may cause spectrum prices and network costs to increase for the larger companies, impairing their ability to serve customers economically. Impeding new entrants (through price or otherwise) from acquiring the spectrum they need to get into the market may limit competition in the long term.
Other issues plague the efficient allocation and use of spectrum. Unlike in the US, there are few functioning secondary markets in Europe, the Middle East and Africa, which means operators may not be able to optimize their holdings through sales, acquisitions or trades with other spectrum holders. This limitation has very real technology and complexity costs.
Ideally, technology licenses should be technology neutral; however, some are technology specific. For example, they dictate that 2G must be used in a particular band, rather than more advanced – and more efficient – 3G or LTE. Other licenses cover bands that are too narrow to be useful or carry timeframes that are too short to justify further investment. Spectrum is often released in small blocks – in some extreme cases as low as 1 MHz bands – which provides limited flexibility and raises costs for operators. At the same time, some operators have yet to build out infrastructure for spectrum they have acquired, turning a scarce resource into a wasted one.
(Lack of) Harmonization
Lack of harmonization at regional and international levels – meaning, for example, that the same operator’s 3G or 4G network operates on different bands of spectrum in different countries or in different regions of the same country – leads to further inefficiency. Currently, for example, 4G networks operate on more than 40 spectrum bands around the world.13 Devices such as smartphones must be designed to work across multiple bands of spectrum, instead of just a few, which is expensive and requires more battery power and antennae complexity. Certain devices are incompatible with particular operators’ networks. Because handset makers focus on the most popular bands (often the bands serving larger markets), smaller operators or operators in some markets may not have access to the most recent devices. New technologies, such as multiband chipsets, are addressing some of these challenges, but lack of spectrum harmonization still imposes inefficiencies and adds costs. Research by The Boston Consulting Group for the GSMA found that countries in the Asia-Pacific region can unlock up to $1 trillion in GDP growth by 2020 through the harmonized adoption of the 700 MHz spectrum band for mobile services.14
Harmonization is a huge challenge because each country has released and allocated spectrum according to its own needs and timing imperatives, and the state of mobile infrastructure development varies widely. No country wants to wait for others to catch up or let others determine the development of its market. Setting out recommendations and procedures to achieve better harmonization at the regional and international level is the goal of the ITU World Radio Conference, which next meets in 2015. There is urgency to this issue. The slow pace of spectrum harmonization processes must be accelerated, or countries that tire of waiting for new processes and procedures will act on their own, further fragmenting an already disjointed system.
Structural Adjustments Are Needed
Many of the problems with spectrum allocation require the efforts of both companies and governments to solve. Governments need to focus on making additional spectrum available while encouraging its efficient use. Companies must make the most of the technology and tools at hand to maximize the capacity of current and future allocations.
Since spectrum is the life-blood of wireless networks, the most important step governments can take is releasing more spectrum for mobile use. This includes traditional licensed spectrum, the top priority for operators in connection with delivering ubiquitous and predictable quality of service, as well as spectrum for new sharing models, including both licensed and unlicensed shared access.
Governments should consider refining auction processes for licensed spectrum. Among the ideas receiving consideration are auctions geared to longer-term value, which charge fees over time based on the value generated by usage, rather than set-level upfront payments. Goals include attracting a wide range of bidders – regardless of size – and ensuring that purchasers efficiently utilize the spectrum they buy. As the range of bidders expands, the importance of including build-out requirements in purchase agreements also increases.
Improving the efficient utilization of spectrum allocations can be pursued through additional methods. Governments can minimize the underutilization of a scarce asset and let operators know that they cannot sit on unused spectrum by ordering appropriate build-out requirements for licensed spectrum and authorizing the claw-back of designated bands if these obligations are not met. Governments can also support the development of secondary markets so that operators have additional opportunities, beyond one-time auctions, to match spectrum acquisitions with their needs.
Other, more technically oriented spectrum management innovations can not only improve use of existing bands, but also enable more rapid incorporation of technological advances that provide efficiency benefits. Regulators should release larger contiguous bands of spectrum that provide operators more flexibility and greater throughput, though new technologies may be reducing this need. When band assignments are directly tied to specific technologies, such as 2G or 3G, utilization can wane as the market shifts to newer and more efficient technologies, such as LTE. Regulators can refarm such bands, and in the process, give licensees future flexibility to deploy their choice of new technologies, subject to effective oversight to ensure compatibility with neighbouring allocations.
Several spectrum-sharing models offer the potential to increase utilization through approaches that complement long-term, exclusive-use licenses. Licensed Shared Access and Authorized Shared Access seek to make broader use of dedicated spectrum that is currently used only at certain times or in particular locations (such as for testing of military equipment, or ship-to-shore radar). These approaches increase efficiency by allowing commercial users to share access on a designated basis, which helps provide the reliability and predictability that operators desire. Unlicensed dynamic shared access models can also work through specific technical rules.
Unlicensed spectrum, also has an important role to play. The best-known unlicensed technology is Wi-Fi, which is now available on billions of devices, and has emerged as an important resource for operators to offload burgeoning data traffic. This will only increase with LTE Advanced technology, which can involve aggregating unlicensed and licensed spectrum in the same network with the same wireless technology. This helps operators augment the capacity of their networks by using the unlicensed spectrum more efficiently while providing a tight interworking between the licensed and unlicensed bands.
As the Internet of Things and M2M services evolve, an ever-broader variety of spectrum needs will need to be filled. New M2M services, such as smart electricity metering, may initially have needs more akin to low bitrate 2G services. However, as M2M communications become smarter – think self-driving cars – more advanced 3G or LTE spectrum may be required. Although operators can likely leverage 2G networks for M2M now, in the long run, as services become more intelligent, they may find it difficult to justify maintaining these networks and using valuable spectrum to support such low bitrate use. Operators should be allowed to recognize the specific needs of these services – and how these needs will evolve – and use spectrum to serve them in an efficient manner.
Unlicensed spectrum will also play an important role in future of M2M. Today, many sensors and M2M services already communicate through Wi-Fi, Bluetooth, radio frequency identification (RFID) and other unlicensed technologies. Satisfying these diverse needs will require balanced policies that provide not only more licensed and unlicensed spectrum allocations, but also more flexibility for shared access to allocate underutilized spectrum.
Operators also need to invest in strategies that maximize the efficiency of their current holdings and the capacity of their networks. Initiatives such as those described above are key. So are incentives for users to move up to more efficient 3G and 4G networks (the majority of the world’s wireless customers, more than 4 billion connections, are still on 2G networks) and to build denser networks with smaller cell deployment models that can handle higher traffic.
Smaller cells will represent a vital, complementary tool for improving efficiency, especially in densely populated areas. While traditional cellular deployment, which relies on relatively few high-powered radios usually mounted on cell towers, has been cost effective, the growing number of users and exploding amount of data are pushing the limits of capacity. By contrast, small cells can be placed almost anywhere – on buildings, streetlamps and bus stops, for example. They can handle a much higher volume of traffic and are adding much-needed density to cellular networks, bringing connections closer to end-users and blurring the distinctions between wired and wireless networks. Mobile networks in Tokyo have already moved towards a small cell approach, with stations spaced every 100-200 meters.15 This is approximately five times the density of a typical urban market.16 One potential model even involves the deployment of open-access small cells in existing premises, such as homes and offices, simultaneously freeing up macro network capacity for other users and increasing network capacity inside buildings, where most wireless broadband data is consumed.
Small cells are a key component of the more heterogeneous network environments – combining macro cells, Wi-Fi and small cells – that are expected to evolve in the near to medium term. Governments can encourage new wireless facilities deployment by quickening permitting and other approval processes.
As a final point, application developers can take spectrum use into account in the technical specifications of their innovations, reducing the spectrum stress that growing volumes of content place on mobile networks. Facebook founder Mark Zuckerberg has advocated making applications more efficient as a means of expanding internet access, particularly in emerging markets, by reducing data delivery costs.