The post-3G wars are well under way. The telecoms industry has set its sights on mobile broadband, which will be a $400 billion global market by 2012, according to Informa Telecoms & Media. And proprietary or market-specific technologies aside - or unless TDD suddenly becomes sexy outside of China - it's essentially a three-horse race between GSM, CDMA and Wimax.
Which horses are competing, exactly, is a matter of conjecture. On the GSM side, it's HSPA and the upcoming LTE (Long Term Evolution) due for commercial deployment in 2010. From the CDMA camp it's 1x EV-DO Rev B and its successor UMB (Ultra Mobile Broadband) slated in 2009. And for Wimax there's 802.16e and, also around 2009, 802.16m.
The sales pitch for each camp can essentially be boiled down to the promise of 'true' mobile broadband connectivity - multi-megabit wireless data speeds anywhere you like without being tied down to coffee shops and airport lounges, and all of it employing the same base technologies of OFDM and MIMO. Yet each camp claims major advantages over the other - i the case of the GSM and CDMA camps, it's the usual economies of scale and interoperability vs technological genius and spectral efficiency, respectively.
Where Wimax stands in all this depends in part on which cellular evolutionary element you're comparing it to. For example, Wimax proponents maintain that mobile Wimax, which is coming out now and will roll out in earnest starting next year, is beating UMB to the punch by at least a year and LTE by at least two or three years.
'If there's not a good fixed-line infrastructure, if you have to move today, the best technology is Wimax, because it performs better than 3G currently does now,' says Anthony Berkeley, director of LTE marketing and strategy for Alcatel-Lucent's wireless business group.
On the other hand, the GSM Association states firmly that mobile broadband isn't future-tech - it's here now with HSPA rollouts. In Asia alone, according to the GSMA's Wireless Intelligence analyst venture with Ovum, W-CDMA connections in Asia will hit 160 million in 2010, 50% of which will be HSPA. The GSMA also says that the current HSPA peak speed of 7.2 Mbps is more than enough for most mobile apps - even data-intensive activities like downloading PowerPoints, video clips and music files.
That said, CDMA vendors like Alcatel-Lucent, Huawei, Motorola, Nortel Networks and ZTE are pitching UMB as 'true mobile broadband', with peak speeds of close to 290 Mbps.
'That's enough bandwidth to do high-definition video over mobile,' says Kimberly Flowers, director of CDMA data product management at Motorola. 'HSDPA and EV-DO Rev A can't do that.'
Either way, the push from the cellular sector to promote their post-3G strategies has re-ignited the questions over the role of Wimax in mobile broadband - or if it will even have one.
'By 2010 we see a potential base of 800 million users for mobile broadband - either LTE or UMB,' says CW Cheung, APAC consulting director at Ovum. 'By 2011 Wimax will account for only 2% of the mobile broadband market.'
To be sure, the spectacle of cellular champions writing off Wimax as an also-ran is nothing new. Meanwhile, Wimax has been making serious headway as a fixed-line substitute in developing markets and generating headlines with big-name cellcos like Sprint rolling out mobile Wimax services in the US and Vodafone trialing it.
Indeed, the whole Wimax vs 3G thing should be old hat by now, especially since vendors that sell to both sides have maintained a 'horses for courses' philosophy - it's not about one technology conquering the other, it's about the business case and deploying whatever technology helps you realize it.
The catch is that Wimax's business case, which was designed to compete with current 3G technology, may have trouble holding its own against the likes of LTE and UMB. Put another way, for some observers, mobile Wimax is in greater danger than ever of being a niche technology - or worse, an interim technology.
First off the blocks
When looking at the forecasts, it's helpful to keep in mind that when it comes to LTE and UMB, there's little mystery as to who will adopt what. W-CDMA operators and EV-DO operators will, given enough time, follow their designated evolutionary paths.
As for mobile Wimax, the conventional wisdom is that it's a natural for players with no 3G license (i.e. fixed-line operators, ISPs, 2G cellcos that lost or ignored their local 3G spectrum auction/beauty contest, etc) or greenfields.
On the other hand, Wimax proponents insist there's a compelling case for cellcos to add Wimax to their service portfolio as well. And some big-name mobile operators have turned heads by embracing Wimax. Most famously, US cellco Sprint is rolling out Wimax service in two US cities this year, which will be extended to 100 million people by the end of 2008.
Admittedly, Sprint may be a unique case, because it gained a wide patch of spectrum when it merged with Nextel in 2005, and the US government in essence said to Sprint, 'use it or lose it.' Sprint's official business rationale for rolling out Wimax is that the Nextel spectrum gives it an opportunity to roll out a service - nomadic mobile broadband in the multi-megabit range - that its competitors can't duplicate (at least not as cost-effectively or as quickly).
Telecom Italia, whose mobile arm offers 3G, is trialing Wimax with an eye on launching nomadic broadband before its competitors do.
'We can't afford to leave Wimax for someone else to be the first mover against us,' says Valentina Riviora, Telecom Italia's project manager for service creation and application.
Then there's Vodafone, which has rolled out a fixed Wimax network in Malta (based on Airspan's HiperMax base stations, which are software-upgradeable to mobile Wimax) and, much to the joy of Wimax backers, became a principal member of the Wimax Forum.
Steve Pusey, global CTO at Vodafone, said, 'Vodafone remains technology neutral as far as our future network options are concerned, and joining the Wimax Forum is a logical step as we seek to evaluate the full capabilities and potential of this technology.' Succinctly put: Vodafone is technology neutral and is interested in Wimax as well as LTE as part of its next-generation network strategy.
Which wouldn't be a big deal, says Ozgur Aytar, an analyst with Pyramid Research, except that Vodafone just happens to be one of the biggest GSM players in the world.
'Vodafone, owing to its scale, is an agent of change in the operator community, and we expect others to follow its trajectory, says Aytar.
Atyar adds that a year ago, a GSM operator adopting a multi-technology network strategy was unheard of. Many LTE vendors assumed that the GSM operators were exclusively committed to LTE, but that's no longer a given.
'With operators present in different markets and looking for new revenue sources, there is no 'one-size-fits-all' technology, but the wrong technology can set an operator back years,' he says. 'In catering to the needs of different markets and customer segments, operator networks will comprise diverse access technologies, each optimized for certain geographies, demographics and services. For the Wimax champions this is good news; for LTE backers it is a strong warning that should lead to increased R&D budgets for the next few years - in both camps.'
Prashkant Gokarn, partner with Spectrum Strategy Consultants, puts it in even stronger terms, pointing out that just because LTE will be available by 2010 doesn't mean HSPA operators will adopt it en masse.
'In many markets 3G makes up only about 20% of users maximum, so what's the rush to invest in LTE‾' he says, adding that by some analysts' reckoning, some markets may see no urgent need for LTE until 2025.
Wimax now, LTE later
The GSM Association's response to such predictions is to point to the rate of growth of HSPA alone, which has gone commercial in over 120 networks in less than two years - it took vanilla W-CDMA almost five years to hit the same target.
'If you look at how fast HSPA is rolling out, we're going to have more 3G subscribers than 2G in less than three years,' says GSMA chair Craig Ehrlich.
For that matter, Ehrlich is not overly impressed with Vodafone's Wimax dabbling either, admitting that Vodafone will naturally try out different technologies, but noting that Vodafone chief Arun Sarin said at the GSMA's annual conference in Barcelona earlier this year that Wimax was a wake-up call to the GSM camp to not settle for HSPA and get cracking on LTE development.
'If we stopped at HSPA, then sure, we'd fall off a cliff and mobile Wimax would have a market. But we're not,' Ehrlich says.
In any case, the notion of rolling out Wimax ahead of LTE raises a valid question for cellcos: is Wimax a stop-gap solution‾ And if it's not, what happens to the Wimax network once you've gone to LTE‾
Alcatel-Lucent's Berkeley insists that there's no reason to look at Wimax and cellular as being mutually exclusive because there is a market for both static nomadic broadband and mobile broadband.
'It's not a question of one technology totally replacing another,' he says. 'If you build a Wimax network for static and nomadic services, you can change it to perform better by cutting off the mobility aspect and make it better as a static service. LTE is designed for high mobility, but Wimax can still serve a specific consumer segment, so there's no reason to scrap it.
Transitional advantage
However, phased-deployment is one element of 3G that doesn't necessarily work against it. For both GSM and CDMA, backward compatibility is built-in, permitting service providers to selectively deploy the next generation of services into the market. Cellcos can start with LTE or UMB high-density urban areas and strategic indoor locations such as airports and business parks (much as they did with initial rollouts of HSPA and EV-DO Rev A or B, respectively) and let the previous generations handle national coverage.
Consequently, service continuity will be a key service differentiation between 3G and Wimax, says Riviora of Telecom Italia. 'If you're moving and you can't get HSDA, you automatically drop down to 3G, or GPRS/EDGE,' she told Telecom Asia. 'That's something guaranteed in mobile - no matter what, you will always have that data connection.'
Another advantage is this: because LTE will be deployed in ITU recognized frequency bands, LTE will eventually have global roaming capabilities similar to current GSM/3G networks. And because the LTE handsets will be backward compatible with GSM and 3G, even at early stages of LTE deployment LTE subscribers will be able to roam throughout Asia.
Independent Wimax service providers will find that service limitations such as spotty coverage without a fallback capability to a lower-speed service and limited roaming capabilities could limit subscriber growth regardless how good or low-cost Wimax is.
All this is assuming, of course, that Wimax's real challenger will be LTE or UMB, rather than the HSPA or EV-DO Rev B networks already in service.
Qualcomm senior VP Jeff Belk has said for several years now that mobile Wimax doesn't hold a candle to any incarnation of EV-DO. GSMA's Ehrlich is equally bullish about HSPA, arguing that even today the mobile Wimax business case is on thin ice.
'When you look at how fast HSPA is rolling out now, and you look at markets like the UK, where you have five operators doing HSPA at 3.6 or 7.2 Mbps before you've even spent a dollar, and you've got to spend 2 billion pounds just to build your network and compete with them, any investor who's considering mobile Wimax has to ask themselves, where is their market‾ And what can they do differently or better than HSPA‾' says Ehrlich. 'As a businessman, I don't see an answer to that.'
What's the frequency, Wimax
One of the perennial challenges that mobile Wimax faces - like any other wireless technology - is securing spectrum. If you're Sprint, you can buy someone who already owns it (e.g. Nextel). For everyone else, it's not nearly that straightforward.
Wimax proponents often cite the technology's ability to work in a myriad of different frequency bands (700 MHz, 900 MHz 2.3 GHz, 2.5 GHz, 3.4 GHz, 3.6 GHz, 5.4 GHz, 5.8 GHz, and possibly others) as a strength. On the other hand, one of the potential downsides of being able to operate at so many frequencies is that it limits the amount that equipment prices drop and also the potential for interoperability. And in the short term, it also means that operators using frequencies other than the initial certifiable bands may not have access to true Wimax equipment for some time.
Of all the potential Wimax frequency bands, 2.5 GHz and 3.5 GHz have held the most promise for a somewhat harmonious approach for Wimax worldwide. However, satellite operators have opposed allocating 3.5-GHz bands to Wimax because of irresolvable interference problems with extended C-band transmissions.
The 2.5-GHz bands are more desirable from a equipment cost point of view, but the band has also been designated an extension band for 3G and will be allocated next year. As such, 3G operators in Europe feel that they already own claim to it in principle, if not in fact.
Even the lower bands are being contended. The GSMA has been calling on country regulators to open up the 900-MHz spectrum band to 3G, which Ovum says could lead to an additional 150 million people across the Asia Pacific - particularly those in emerging markets - to go 3G by 2012 by providing better rural and in-building coverage.
Though the GSMA makes no mention of Wimax in its 900-MHz brief, such a move would hinder (if not block) Wimax's growth in these markets, as well as provide spectrum that could later be re-farmed to LTE.
- Chris Everett
Mobile broadband is only skin-deep
UMB claims a peak data rate of 288 Mbps; LTE claims 100 Mbps; and Wimax chimes in with 46 Mbps, but if you will wait a little while, Wimax will jump to 1 Gbps with the introduction of 802.16m. So after the dust settles what can be said‾ First, all three systems will use the same air interface technology, OFDMA. Second, all three systems will use the same antenna technology, MIMO. Third, all three systems are banging their heads against the ceiling known as Shannon's Law.
Shannon theorized that the amount of error-free data that could be transmitted over a channel of any given bandwidth is limited by noise. For example, assuming that the thermal background noise in the network produces a signal-to-noise ratio of 20 dB and that there is no other sources of interference in the network, then Shannon's Law predicts that no more than about 6 bps per Hz of error-free information can be sent.
LTE is targeting a 100-Mbps data rate on a 20-MHz bandwidth channel. That equates to 5 bps per Hz, which is near the theoretical limit.
However, the real world consists of more than thermal noise. There's noise generated by other users, from the cell site, and from neighboring cell sites. All that interference lowers the signal-to-noise, reducing both the theoretical (Shannon's) and real-world capacity available over any channel.
Therefore, all three systems are facing the same limiting capability of how much error-free data they can place on a unit of bandwidth. To increase data rates, each system must increase the bandwidth used to transmit data to a user. HSPA claims a peak data rate of 42 Mbps on a 10-MHz channel; LTE claims 100 Mbps on a 20-MHz channel. Most of LTE's increased performance is based on increased channel bandwidth, not on a change of technology from W-CDMA to OFDMA.
Thus, a major key as to how successful each technology is in delivering the most data to the most users at any one time is how well each system manages its bandwidth allocation to each user.
And since all three systems are banging their heads against the same ceiling, perhaps real-life data rates won't be all that different.
- Chris Everett