Spectrum policies: a delicate balance

It is likely that a combination of wired and wireless technologies will be required to meet ‘broadband for all’ ambitions. There is an expectation that further spectrum will be needed for mobile broadband systems, over and above that already allocated and in use.
A number of governments have published their views on how much additional spectrum might be needed – for example, the US government announced in its National Broadband Plan that in the next decade it intended to find another 500MHz of spectrum for mobile services. It has been estimated that the UK will require a similar amount of extra spectrum.
At an international level, the ITU Radiocommunication Sector (ITU-R) is responsible for reviewing frequency allocations every few years at World Radiocommunication Conferences (WRCs). Additional spectrum for mobile services is not on the agenda for the forthcoming WRC in 2012 (WRC-12), but it is expected that it will be considered at the following one (WRC-15).
Estimating the future spectrum requirements for mobile services is a complex issue, requiring consideration not only of the expected growth in use of mobile data services during the next decade, but also of various economic and market factors to determine the peak traffic load that mobile networks might need to accommodate.
International Mobile Telecommunications (IMT) is the name given to the family of 3G and 4G mobile standards defined by the ITU-R. IMT-Advanced describes the evolution of the original 3G standards towards 4G. Spectrum requirements for IMT systems were last considered by the ITU at the World Radio Conference in 2007 (WRC-07), with estimates of spectrum requirements based on a methodology defined by ITU-R Working Party 5D, which is part of ITU-R Study Group 5. But the significant increase in use of mobile broadband services means that forthcoming WRCs may need to revisit spectrum requirements. Another new methodology will likely need to be found to estimate future spectrum requirements.
Around 10 years ago, the ITU-R developed a spectrum calculation methodology (ITU-R M.1390) to estimate the amount of spectrum that IMT systems required, both then and in the future. The methodology was based on estimating system capacity, assuming traffic was carried over circuit-switched networks – typical of 2G mobile systems.
For WRC-07, a new recommendation was made (ITU-R M.1768), which defined a methodology for calculating spectrum requirements for voice and data services. This methodology is applicable to both circuit-switched and packet-switched traffic, and still provides a valid means of estimating total spectrum requirements.
However, we believe that future estimation models need to take account of various interrelated market and economic factors relating to 3G and 4G traffic types and loading which, taken together, will ultimately determine the economic traffic load per mobile network.
Factors behind traffic demand
Various supply-side factors influence the market for mobile data services, which ultimately determine the supply curve, and therefore the cost per bit for mobile networks to carry data traffic of various types.
These include: improvements to network technology, such as evolution from HSPA to HSPA+ and/or LTE, and the improved capabilities this provides to carry data traffic; use of small-cell technologies to improve capacity; the potential for innovative spectrum use, for example, cognitive techniques or use of ‘white spaces’; and the effective availability of spectrum to the market, which is determined by national regulatory decisions to refarm or release new spectrum to the market.
Demand-side factors include various issues that in some cases lead to an increase in demand or traffic load (for example, improved mobile device design driving greater usage, or availability of new services), and in other cases to a reduction in demand or traffic load (for example, as a result of offloading traffic to Wi-Fi or femtocell networks, or improvements to fixed networks). Operators’ pricing strategies also have a significant impact on the take-up of mobile data services.
New estimation models should therefore forecast the spectrum required to meet only the traffic demand that can be economically delivered via mobile networks – rather than the total mobile data traffic, which might be carried over a combination of mobile and other networks.
This suggests that future models must consider:
· how the demand for different mobile data services may vary according to price and operator pricing policies
·  the cost incurred by mobile operators to meet the future traffic demand
· the availability of other wireless networks, such as Wi-Fi and potentially other hotspot networks, to meet the demand that cannot be profitably carried by mobile networks.
It is also noted that solutions to meeting the future spectrum requirements for mobile broadband systems might need to consider requirements to redistribute spectrum from other uses. But redistribution of spectrum on its own is unlikely to meet future needs, so a combination of solutions - involving additional spectrum, more-intensive use of existing spectrum, and use more spectrally efficient technology - is likely to be required.
Ultimately, being able to match future demands for spectrum with available resources relies on a detailed understanding of the distribution and use of spectrum to be able to make choices on future use. The creation of spectrum inventories is one way to achieve this – as proposed and implemented by regulators in various countries in recent years.
Janette Stewart is a Senior Manager at Analysys Mason. For more information, go to www.analysysmason.com/