The battle for supremacy between the emerging LTE and WIMAX 4Gmobile standards isn't over yet.
Casual observers of the telecoms sector might be forgiven for thinking that the battle to become the global standard for next-generation (or 4G) mobile networks had already been won, with LTE easily beating WiMax. The reality isn't that simple.
Each camp has been busy developing its technical specifications and lobbying the network operators, governments and equipment vendors who will ultimately decide who wins.
While a number of wireless broadband services and devices (including mobile phones, USB dongles and laptops) are already being marketed as '4G', the world is yet to witness the first official transmission and reception of an IP data packet over a proper fourth-generation cellular network.
The body that decides what is or isn't a 4G network is the International Telecommunication Union. ITU-R, which is in charge of radio communications, sets out the criteria for a 4G radio interface (or, more technically, as compliant with the International Mobile Telecommunications-Advanced system).
IMT-Advanced specifies, among other parameters, the use of OFDM (orthogonal frequency-division multiplexing) digital modulation, all-IP packet switching, and peak download data throughputs of at least 1Gbit/s and 100Mbit/s, respectively, for low- and high-mobility scenarios.
Neither WiMax nor LTE support these throughputs today. The situation is a repeat of what happened ten years ago, when the GPRS and EDGE mobile-phone standards were introduced to offer greater functionality than their 2G (GSM/CDMA) predecessors. Unfortunately, their performance was not enough to be recognised as 3G by the ITU-R; so they were labelled as 2.5G.
Similarly, the first two versions of LTE (3GPP Release 8 and Release 9) are a step up from so-called 3.5G (HSPA) and even 3.75G (HSPA+), but don't quite make it to full 4G compliance. So the wireless industry is calling LTE, somewhat ludicrously, a 3.9G air interface.
The WiMax standard has much shallower roots. There has never been a 2G or 3G WiMax technology. In fact, the first versions of the specification (collectively known as IEEE 802.16) were designed not for mobile, but for fixed wireless broadband access.
When the 'mobile WiMax' (802.16e) version of the standard was released in 2005, it appeared clear that this radio interface was on a collision course with the established family of cellular platforms.
Such suspicions have been heightened recently by the realisation that the only two radio interface standards that will be submitted to the ITU-R to be recognised as 4G are LTE-Advanced (3GPP Release 10, scheduled for completion in 2011) and WiMax 2 (IEEE 802.16m, also expected to be finalised next year).
Until recently, the consensus among industry watchers was that a head-on collision between the two standards was unlikely. This was because they catered to the needs of different wireless service providers.
Incumbent mobile phone operators that had grown up with the evolving GSM radio standard were thought likely to continue down that path - what with the advantage of its access to a rich ecosystem of basestations and user devices that kept equipment costs down and ensured backwards compatibility and guaranteed roaming connections.
WiMax, on the other hand, arrived with the promise of bringing economies of scale to new-entrant, greenfield operators for whom - because of their lack of financial muscle, the terms of their licences or the amount or type of spectrum that they owned - LTE was not an option.
Following this logic, alternative wireless operators began building WiMax networks and launching commercial services in 2005 and 2006. Today, there are nearly 600 WiMax networks deployed in 149 countries, operating mainly in the 2.3GHz, 2.5GHz and 3.5GHz frequency bands.
Declan Byrne, the WiMax Forum's director of communications, says WiMax antennas are now covering a combined area worldwide that would allow more than 650 million people to subscribe to a service if they chose to. 'That number is expected to grow to something like 800 million by the end of 2010 and surpass a billion in 2011,' he says.
But 'potential subscribers' is a very different measure than 'actual subscribers'. Of the latter, Byrne admits there are 'just over 10 million around the world right now'.
Richard Webb, directing analyst, WiMax, microwave and mobile devices at Infonetics Research, contrasts that figure with the more than 200 million global HSPA subscribers that cellular operators have amassed in a similar period. 'LTE subscribers are forecast to be far in excess of that,' he says, 'so WiMax really looks like the baby brother.'
Webb adds that no country where WiMax has been deployed is mustering huge numbers of subscribers at the moment: 'Most countries have under a million, which doesn't look particularly impressive.'
By far the largest WiMax deployment to date (in terms of both coverage and number of subscribers) is that of Clearwire in the US, under the Sprint brand. A few weeks ago the operator said it now had 1.7 million retail and wholesale subscribers, which is 231 per cent more than 12 months ago - and only half of what it expects to have by the end of the year.
Together with Clearwire, two other wireless broadband ISPs (Yota in Russia and UQ Communications in Japan) are usually singled out as the three 'poster children' for large-scale WiMax rollouts. Earlier this year, two of these operators announced plans either to deploy LTE instead of WiMax in some of the cities to which they will be extending service (in the case of Yota) or to try LTE both as a standalone and as a coexisting WiMax air interface (in the case of Clearwire).
Unsurprisingly, the announcements were widely interpreted as major blows to the WiMax camp. 'One year ago I would have seen no reason for a WiMax operator to want to move to LTE,' says Monica Paolini, president of Senza Fili Consulting. 'At that point, LTE was mostly considered to be an FDD [frequency-division duplexing] technology,' while most greenfield operators prefer - or have no choice but to use - time-division duplexing (TDD).
But things can change fast. A'TDD version of LTE (TD-LTE), developed for the Chinese market, has suddenly started to becomevery popular among mobile operators elsewhere. 'This gave WiMax operators the opportunity to adopt a more widespread technology in their networks,' says Paolini.
'This has become an option for a lot of them, and it's interesting because, from a performance or a [basestation] cost point of view, both air interfaces are pretty much the same. There are some differences but not order-of-magnitude differences. However, with TD-LTE operators do have the ability to tap into a larger market for devices, which means more choices and lower costs. Also, they're going to find it easier to roam with other operators or to allow visitors to use their network,' Paolini believes.
It is a view shared by Webb: 'When you've got WiMax-friendly spectrum that is now also usable for TD-LTE, then you've a got genuine technology decision to make as an operator, and it does become a head-to-head trade-off.
'At that point, an operator weighing the decision has to evaluate the devices ecosystem, which - in the case of WiMax - is pretty limited and pretty modest (but it's available now), versus the expected, much bigger ecosystem for TD-LTE (of which none is available now).'
A picture seems to be emerging of an industry at a crossroads. Rupert Baines, vice president of marketing at picoChip, whose silicon designs are used by both WiMax and LTE modem makers, says: 'There's a lot of commentary but, at the moment at least, we're not really seeing any of the operators [being linked with a potential migration towards LTE, such as] Yota, UQ or Clearwire halting investment in their WiMax networks. And my gut feeling is they will carry on investing in them.'
Ongoing investment in WiMax infrastructure is, rather than a gut feeling, a dead certainty if you already have a working network to maintain and paying customers to keep happy.
But then the questions start. At what point do you switch WiMax off once the LTE ecosystem begins to flourish? At what cost? And how technically viable is it to migrate smoothly enough so that users won't lose faith in the service?
'People do change technology tracks,' points out Baines. 'AT&T changed from TDMA to GSM - it was expensive and painful, but they had to do it.'
The other big question is: what do you do if you are a greenfield operator that has a licence to exploit a chunk of spectrum for which either WiMax or TD-LTE would be suitable and you haven't started building your network yet?
It's a question for which we might soon have an answer, as this is the situation in which a group of would-be wireless broadband operators in India finds itself. Following an intensive bidding process, Reliance, Aircel and Airtel are among the highest-profile operators that now have to decide which air interface they choose to go with their new 2.3GHz spectrum holdings.
A source interviewed for this article who asked not to be named put forward an interesting argument in favour of WiMax: 'If you're Reliance and you've just spent $2.5bn on spectrum for India's sole nationwide licence, the interest charge you'd get on that alone would be over $1m a day.' Acting quickly would clearly pay off, at least in the short term.
Still, rumours have surfaced that Reliance is getting ready to go with TD-LTE. But the WiMax Forum holds out hope. 'We were a bit distressed to read that one of the first utterances from Reliance Industries was that they were very attracted to the future ecosystem of LTE,' says the WiMax Forum's Byrne.
'We have met with Reliance at the most senior of levels since the auction closed and there were at least two instances where they talked publicly about their attraction to WiMax as a ready-to-deploy technology.'
'I don't know what's going to happen,' says picoChip's Baines, 'but if WiMax has 50 million-plus users worldwide in 24 months, then it will no longer be considered niche, but mainstream.'
According to Paolini, in 24 months there will actually be more that 60 million mobile WiMax subscribers around the world. 'There are several new WiMax deployments taking place, while a lot of existing deployments are being extended,' she says. 'There is a lot of activity. Having said that, a lot of those operators are also considering moving to LTE eventually.'
In Webb's opinion, WiMax is only going to succeed if it finds complementary markets to LTE. 'If it goes head-to-head to LTE it won't win that battle.'