South Africa Goes Its Own Way at 2.6GHz
ICASA has now finally announced or confirmed its spectrum plans for the 2.6GHz band in which four licenses of 30MHz will be awarded. This structure puts South Africa firmly in the camp of “I’ll do it my way” with regard to likely deployments of wireless systems in this band, which otherwise has the potential of being the most widely adopted standardized band for broadband mobile communications services on a global level. The structure of 30MHz unpaired channels reflects the wishes of the advocates of TDD mobile WiMAX technology and their claims that this capacity will enable them to maximize their spectral efficiencies and offer adequate capacities.
Unfortunately ICASA’s decision in this case is far from being in the best interests of the country and its residents. It is highly unlikely to foster the goal of achieving the maximum coverage, minimum costs, and highest quality and performance of future mobile broadband services in the Republic. It is the result of shortsighted thinking which has been unduly swayed by arguments from advocates of one specific wireless system (namely TDD mobile WiMAX), thereby in effect also violating the principle of technology neutrality more than it need be, recognizing that “absolute” technology neutrality is not possible in practice due to the need to manage interference that is inevitable when multiple users, especially if using multiple technologies and/or with varied service requirements, occupy the same and adjacent frequency bands.
The four 30 MHz 2.6GHz licenses planned by ICASA ensure that this band will have a unique structure in South Africa compared to elsewhere. This structure entails significant risks of higher equipment costs, more difficult interference management challenges, and reduced international roaming possibilities for South Africans abroad and for foreign visitors to the Republic. It also makes it likely that significant mobile networks will be deployed in South Africa committed to a technology road map that could soon become a dead end.
Consequences
ICASA’s 2.6GHz spectrum plan entails several unfortunate, but not unlikely consequences:
A dead end roadmap
Assessment of global wireless R&D reveals a very real possibility that the next generation of mobile WiMAX systems (802.16m), which is being designed to offer true 4G services that meet the goals of ITU-Advanced, will never be deployed commercially, i.e. mobile WiMAX, like the 3GPP2’s UMB (UltraMobile Broadband) and earlier the U.S. 2G TDMA standard, will become a dead end road map. The second largest (with about 500,000 customers) mobile WiMAX operator Yota (Russia) has announced its intent to switch to LTE, while the largest one (Clearwire in the U.S.) has been laying the groundwork and creating expectations for doing so at a time of its choosing (which is not likely to be before 2012-13). Equally significantly the preponderant weight of vendors’ and components suppliers’ development work has been shifting in one direction, away from WiMAX towards LTE, notably among large players who had earlier been publicly strongly (even if not exclusively) committed to WiMAX (e.g. Motorola and Alcatel-Lucent, while Nortel has gone bankrupt, and Cisco has withdrawn from the WiMAX radio access network market). If operators who win the 30MHz licenses decide to deploy TD-LTE, then they will have embarked upon the long term LTE road map, in which TD-LTE is being designed to maximize its compatibility with FDD (paired spectrum) LTE so that users can operate seamlessly within network footprints that may contain TD-LTE “hotspots” within FDD LTE regions. However if license holders deploy mobile WiMAX then they (and their customers) may well encounter additional and unnecessary costs and disruptions (even though WiMAX and LTE have many technological similarities) when in future they have to migrate to LTE technology, in order to keep up with the improved performance levels of, and demands for, the greater capabilities of next generation technologies.
Limited and more expensive devices
The use of a non-standardized band structure means that the range, time-to-market (despite any early advantage of WiMAX compared to LTE), and costs of the devices available to broadband mobile users exploiting the 2.6GHz band in South Africa will be respectively smaller, longer, and higher compared to the benefits of economies of scale and development priorities of device and components vendors with respect to commercializing devices, notably HSPA- and LTE-capable ones, that can address much larger markets than South Africa alone represents. Furthermore, LTE is being commercialized by an impressive global group of powerful competitors who give each other no quarter. So the argument that it is dangerous to rely on one next generation technology, because this dependence may inhibit the growth of a highly competitive supply situation for the benefit of users and operators, is not credible, as the example of Ethernet shows in another networking context.
Interference and Roaming
The deployment of 2.6GHz networks in non-standard and perhaps even unique band structures could introduce additional challenges of interference mitigation in border regions with countries which adopt more common structures (e.g. ITU Option 1), so that networks may be deployed in paired operation at the same frequencies as they operate in unpaired mode on the other side of the border. The consequences can be some loss of coverage and capacity. The severity and impact of these consequences are greatest in parts of the world, such as Europe, where border regions can cover a significant proportion of the total area of a country and population densities in these regions are high (e.g. Belgium, Netherlands, Germany, France etc.).
Another consequence of a highly non-standard 2.6GHz band structure is that it reduces the scope of future international roaming services that would otherwise be available for South Africans in countries that have adopted and are likely to adopt a standardized 2.6GHz band structure, e.g. Europe (and for Europeans in South Africa).
Other Comments
ICASA’s 2.6GHz band plan much more obviously violates the principle of technology neutrality than other schemes. I have pointed out elsewhere that “pure” or “absolute” technology neutrality for spectrum use is not possible. Regulators and spectrum management policies also have to strive for the goal of ensuring that interference between multiple operators in a band, and if such is the case multiple wireless technologies, can be managed at minimum costs, loss of capacity, and risk of adverse impact upon service quality for the sake of all users. This goal inevitably leads to decisions about frequency allocations, guard bands, and power levels that are likely to favor one or some wireless systems over others. ICASA’s decision to allocate unpaired 30MHz licenses exhibits a manifest bias in favor of mobile WiMAX, which is for now only available for TDD operation. TD-LTE for example can be deployed in channels from 1.4 to 20 MHz, while by definition FDD LTE – and FDD WiMAX which is being developed – is being designed to work in paired channels up to 2x20 MHz. The choice of 30MHz channels for unpaired operation in the 2.6GHz band, which reflects lobbying by WiMAX interests, also means that if more than one such license is awarded it becomes problematic to adjust subsequently and easily to the preferred ITU Option 1 structure for 2.6GHz which incorporates 50 MHz of unpaired spectrum in its mid-band, and 140MHz of paired spectrum in its upper and lower bands.
Short Sighted Perspective
Since any band structure, as it must, tends to favor, but not necessarily decisively, the deployment of one or some wireless systems over others at one moment in time, it makes sense to include a consideration of evidence relevant to the likely long term road maps of these various technologies, which may change their respective outlooks. Licenses and their impact upon markets typically last longer than the cycle for the introduction of new, much more powerful versions of technologies. In the context of the substantial and growing body of evidence regarding the risky prospects for mobile WiMAX compared to LTE beyond the next few years, ICASA’s spectrum plan for 2.6GHz appears to be an example of very short sighted thinking. Its 2.6GHz spectrum plan does not serve the best interests of the country or its mobile broadband users and operators, that lie in facilitating the deployment of increasingly powerful and affordable mobile broadband networks and services over the next decade and beyond. ICASA seems not to have taken adequate account of global commercial and operational realities or to have applied a sufficiently long term perspective in formulating its 2.6GHz spectrum plan. The adverse consequences of this decision will be felt for a long time.