WiMAX and LTE butt heads in the ITU and the Indian auctions

WiMAX finally achieved this in the 3G bands when it was accepted by the ITU as an IMT-2000 or 3G approved technology - too late for the first wave of 3G build-outs in advanced economies, but making it possible to bid for a role in the second wave, such as Europe's 2.5-2.6GHz plans, and for emerging economy contracts. Now it will need to gain the same recognition in IMT-Advanced, in order to achieve the essential first step of eligibility to bid in bands usually associated with mobile operators. Nowhere are the conflicts this issue throws up more apparent than in India, where regulator TRAI is facing the latest in a line of contentious spectrum decisions, this time regarding the bands where 3G and 4G will live in the world's fastest growing mobile market.

In targeting approval for the IMT-Advanced program, both the IEEE 802.16 working group - which is looking at the next iteration of the standard that is the basis of WiMAX - and the 3GPP, are seeking to define their next generation platforms, in the hope these will meet the ITU requirements for 4G. This may seem premature - those requirements are not finalized; the first version of the LTE standard is also unfinished; while the WiMAX community is still waiting for volume certification of the all-important wave 2 of the Mobile WiMAX system. However, positioning is critical, and even just-on-paper standards may help influence the ITU's thinking.

Certainly, current 802.16e and LTE platforms come nowhere close to the assumed definitions of 'true 4G', including 100Mbps mobile speeds and 1Gbps fixed rate, although both incorporate certain key technologies that will be specified by the ITU, such as MIMO antenna arrays. But the next generation developments need to be accelerated, if only to gain ITU and operator attention - and it is at this stage that some carriers and other interested parties believe WiMAX and LTE may converge, with 802.16 forming the basis of the TDD element of a next wave standard. The main fly in the ointment of this apparently neat vision - as expounded repeatedly by Vodafone CEO Arun Sarin - is that there is already a TDD technology for LTE, mainly based on work done in China, whose 3G and 4G platforms have been TDD-based. Ericsson, in particular, has recently been working closely with Chinese parties to bring the country's technology into the LTE mainstream, and the TD-SCDMA Alliance, the industry body behind the Chinese 3G standard, announced last week that the LTE iteration would be submitted to the ITU by the end of this year for inclusion on its IMT-Advanced list. Manufacturer Putian says it will have a prototype TD-SCDMA LTE device ready in the fourth quarter, delivering peak data rates of 100Mbps while fully mobile. With most 4G carrier deployments likely to be focused on broadband data services, the importance of the TDD aspect of the standards will be far more important than in previous generations, when most services were voice-oriented and so most spectrum allocated was FDD.

Over in the LTE mainstream, the 3GPP has started work on LTE-Advanced, and plans to firm up the basic specifications at the body's next meeting in Prague, Czech Republic at the end of May. NTT DoCoMo and vendors such as Nokia Siemens have already been key contributors and shown demonstration systems that far outperform current WiMAX or LTE, and point the way towards the extreme spectral efficiency required for 'true 4G'.

The IEEE released its draft description of the 802.16m standard, the successor to 802.16e, earlier this year. The working group is headed by Roger Marks, also a senior VP at NextWave Wireless, and Marks claims 16m will be approved and frozen by the end of 2009, which he claims would give WiMAX a headstart of 18 months over LTE-Advanced.

The actual technical requirements and evaluation criteria for IMT-Advanced should be decided at a meeting of the ITU's radio working group in Dubai at the end of June, and the deadline for proposals should then be October 2009.

While the official standards processes grind on, the operators have more immediate decisions to make about spectrum and technology choice, nowhere more than in countries where 3G and mobile broadband will be opened up in parallel rather than in sequence, as looks likely in India. The Indian regulator TRAI aims to auction 3G bands this year (in India these will be 450MHz, 800MHz and 2.1GHz), but also spectrum in 2.3-2.4GHz, 2.5-2.69GHz, and 3.3-3.6GHz. All these are key bands adopted for Mobile WiMAX profiles, and India is likely to be the largest short to medium term market for the technology, with operators looking to meet fixed broadband targets and leapfrog 3G services in other regions, using a common platform. However, WiMAX' hopes of having these higher frequency bands almost to itself - taking advantage of urgent deadlines of Indian carriers and its headstart over LTE in the 2.xGHz area - may be dimmed by TRAI supporting the 2.xGHz bands for 3G also.

TRAI has launched an industry consultation paper on allocation and pricing for the three WiMAX-targeted bands, and aims to solicit comment from the stakeholders on issues such as eligibility for allocation of spectrum, maximum amount of the spectrum for each bidder, and pricing. At the core of the debate is the different definition of service for the bands, especialy the 2.3-2.4GHz and 2.5-2.69GHz frequencies. TRAI notes that 2.3GHz  was recently approved by the ITU for IMT-2000 or 3G, and 2.5GHz already was, so carriers could use W-CDMA/HSPA  in these bands as well as the official 3G spectrum. It also notes that the 2.xGHz frequencies can support two variants of the 802.16 standard - Mobile WiMAX and the 'OFDMA TDD WMAN' as defined by the ITU. 

"Now with the likely vacation of spectrum in these two bands, it needs to be decided whether the allocation and pricing methodology recommended earlier for the BWA (broadband wireless access) technologies should continue to be applicable for these bands or they should be treated at par with the spectrum bands identified for 3G for pricing and allocation," TRAI's paper says.

This highlights one of the key issues that WiMAX faces as it moves out of its traditional, mainly fixed, broadband wireless access niche into the mobile broadband mainstream, sharing spectrum allocations and target operators with HSPA, CDMA2000 and LTE. BWA spectrum has traditionally been valued far lower than mobile spectrum because of its lower commercial potential, but with mobility becoming more practicable in higher frequencies and all the 4G contenders fighting for the same frequencies, pricing in 2.xGHz is likely to shift towards cellular, rather than BWA norms. This is less likely in 3.5GHz, which remains almost exclusively a WiMAX preserve - its poorer propagation for full mobility and rural/suburban systems has left it ignored by the LTE community, though this may change as techniques to counterbalance those weaknesses are enhanced, and as operators look for spectrum that will support 4G 'hotzones' in high value, urban areas (ie high frequencies and plentiful bandwidth, both promised by 3.5GHz).

Some territories, like the US, will seek to let the market decide, by supporting technology and service neutrality and running auctions. In other economies, the mechanisms of the spectrum and wireless services markets are more controlled and artificial, and in India, where two large operators remain state-owned and TRAI sets detailed rules for auction prices, this is certainly true. However, the decisions TRAI and the Department of Telecommunications (DoT) reach will set some important precedents for valuation of former BWA spectrum, and provide some indicators of how 2.xGHz auctions will be regarded around the world.

According to an earlier TRAI recommendation to the DoT, BWA spectrum should cost only a fraction of the price of 3G spectrum. For example, a block of BWA spectrum in Mumbai has a reserved price of R10Cr (about $2.5m), while a block of 3G spectrum in the same area has reserved price of R80Cr ($19.5m). Further complicating the issue is that the previous TRAI recommendation called for BWA spectrum to be allocated as 15MHz chunks per operator while 3G spectrum was to be assigned in blocks of just 2x5MHz per operator.

This means that carriers could buy BWA spectrum cheaply and then run superior mobile broadband services, using 3G or 4G technologies, to those supported by operators in the narrower, and costlier, 3G-only bands. This, in turn, could reduce interest in bidding for the 3G licenses, and depress their value, or lead to a reduction in real competition.

The obvious solution seems to be to adopt technology neutrality, as in the US, UK and some other European markets going towards 2.6GHz auctions, and TRAI seems to favor this, except for one caveat - that too much focus on mobile broadband could push the price of the 2.xGHz licenses beyond the reach of regional or rural BWA operators, damaging the urgently needed program to improve India's dire broadband penetration record, and could distract providers from basic fixed services in pursuit of mobile media offerings for the growing middle classes.

"However, in view of the compulsion for need for increasing penetration of broadband in the country,  limited availability of spectrum in the identified bands of 2.5-2.69GHz and 2.3-2.4GHz, and more importantly the fact that today no spectrum has been allocated especially for the growth of broadband wireless technologies, one of the views is that, in slight variance with the international practises, why not to consider these bands for use of only BWA technologies for the time being? Based on future developments this can be reviewed," TRAI's paper adds. This would be good news for WiMAX, since it is the only standards-based technology that effectively address the mainly fixed BWA sector (802.16e does this even more efficiently than the fixed-only 802.16d), but it would be a blow to the larger WiMAX vendors that still believe the really lucrative contracts will come with mobile broadband and offering a 4G solution.

The dilemma for WiMAX has always been the same, and is just highlighted by the Indian conundrum - whether to focus on a market it can safely monopolize, such as fixed/nomadic BWA in higher frequencies; or go after the bigger prizes of mobile broadband, taking on the hefty competition from HSPA and LTE for the operator dollars. Unless convergence is attained during the IMT-Advanced process, this question will remain, and in the first years - the two years much touted as WiMAX' headstart over LTE - most of the dollars, outside the US, are being spent on fixed business models.

Note: TRAI will auction five licenses of 5MHz each in the 3G bands, per 'circle' or telecoms region. Although these licenses were initially ruled to be restricted to existing fixed and mobile operators in the country, the government is widely expected to set this condition aside and allow outside bidders, with AT&T and Deutsche Telekom among those rumored to be interested.