One of the notable things about the 2008 Mobile World Congress was the major splash made by Wimax, a technology that the GSM Association has denounced as a niche technology with virtually no future as the world's cellcos inevitably evolve to LTE. Yet there was the Wimax Forum promoting over 40 stands worth of real equipment and boasting over 200 commercial deployments.
Almost as if by response, the 2009 MWC was all about LTE. Numerous heavy-hitters showcased LTE gear in action during the event, with LTE base stations running live video and other bandwidth-heavy apps. Motorola executives drove a van around Placa Espanya running LTE between the van and its booth via two eNode Bs running raw 8-Mbps MPEG-4 video with no buffering or FEC of any kind. The demo went off without a hitch - the handover from one Node B to another produced a bit of picture breakup, but given this was unbuffered video sharing a link with two laptops running YouTube videos, it wasn't bad.
Real-world conditions, of course, won't be nearly as accommodating, but new results released that week from the LTE/SAE Test Initiative (LSTI), which puts an emphasis on real-world conditions in its testing, were encouraging. The LSTI wrapped up its proof-of-concept phase earlier this year, and all up, LTE is essentially living up to the 3GPP's performance benchmarks for downlink/uplink throughput performance and latency.
Put another way, LTE works. It hasn't been vetted for interoperability yet - that's next on the LSTI's docket -but it works.
That was this year's LTE message. It works as advertised, and it's coming sooner than you thought. This year, actually. Various vendors like Ericsson, Motorola, Huawei, ZTE and Alcatel-Lucent intend to have LTE gear commercially available in the second half of this year. And according to Pyramid Research, a dozen operators are committed to launching commercial LTE services sometime next year, including NTT DoCoMo, TeliaSonera and Verizon. Another 15 operators plan to roll out LTE in 2011-2012.
That would seem to be grim news for Wimax, whose chief selling point has been that LTE is years away from commercial reality and thanks to the rise of smartphones and dongles, demand for "real" mobile Internet needs to be met now. The spectre of LTE just around the corner makes that pitch a little harder to sell.
On the other hand, that presumes the market will only accept one or the other, not both. And the market has already accepted Wimax to some degree. The Wimax Forum says the technology had racked up deployments in almost 140 countries (albeit not all of them fully commercial and many of them as wireless DSL rather than fully mobile offerings) as of April this year. Asia has already seen high-profile launches in Malaysia and Taiwan, and will see another in July this year when UQ Communications - a consortium that includes EV-DO operator KDDI - launches commercial service in Japan, where LTE is expected to go live next year.
Moreover, some doubts remain over just how much of a head start Wimax will ultimately have over LTE - and how much of a difference it will make in any case.
Time to market: overrated
Certainly the LTE camp is predictably unimpressed with Wimax's head start in the wireless broadband space, effectively dismissing the "window of opportunity" advantage as irrelevant.
"Time-to-market does not equal success," says Dan Warren, director of technology at the GSM Association, arguing that GSM's history and footprint gives LTE "two equally important things that Wimax lacks -pedigree and a pre-existing ecosystem."
On the other hand, the same is less true for the early-adopter market partly responsible for driving LTE's accelerated development - many of them cellcos who happened to be outside of the typical W-CDMA cellco demographic, and thus have an incentive to push on to 4G sooner rather than later.
"CDMA operators like Verizon and KDDI are in a hurry to deploy LTE because they are stuck on a one-way street with EV-DO," says Bo Ribbing, strategic marketing director for Ericsson. "China Mobile has TD-SCDMA and wants to move quickly to TD-LTE.
And TeliaSonera, which was the first to sign an LTE contract, is sharing its HSPA network with someone else and wants their own network."
For most W-CDMA cellcos, though, LTE isn't an immediate concern, with many currently not planning to deploy LTE for another two or three years. Some are also still considering the interim step of HSPA+, which boosts data speeds to 21 Mbps, although vendors like Motorola counsel against this.
"The window of opportunity for HSPA+ is only one and a half years, so we recommend to operators to go sooner rather than later to LTE because of the benefits of lower TCO, faster time to market and you can better meet mass-market traffic-revenue challenges at a lower cost per bit," says Daisy Lam, senior marketing manager of wireless broadband technologies for the Home & Networks Mobility division of Motorola Asia Pacific.
So far, only a handful of cellcos have adopted HSPA+, but one such cellco - Hong Kong CSL - swears by it.
"HSPA+ has a lot of life in it because everyone is scaling back their capex, and LTE is a costly upgrade," says CSL's chief technology officer Christian Daigneault. "LTE will come, but you need a hell of a lot of spectrum if you're going to achieve the 120 Mbps they're promising."
It's also worth noting that CSL itself has plenty of spectrum to play with - 900 MHz, 1800 MHz (for which it has two licenses), 21. GHz and its newly-awarded 2.6 GHz. And it's using new software-defined radio (SDR) base stations supplied by ZTE to refarm and reuse spectrum across all of its frequency bands.
"The regulatory environment in Hong Kong allows us to refarm spectrum however we like based on market trends and technology, which gives us much more flexible use of our capacity," Daigneault says.
LTE base stations will have the same capability, and that's a potential advantage over Wimax, which so far has typically been allocated single frequency bands, in markets where it operates.
The first delay
However, Verizon - which made a splash in Barcelona by explaining its game plan for LTE, complete with a list of supply contract winners - has already dropped the first bomb on LTE's ambitious timeline. The US carrier originally planned to have its first commercial LTE services up and running in 20 to 30 markets the first half of 2010, and nationwide coverage (in the cities, at least) by the end of 2013. But in a conference call in mid-May, wireless chief Lowell McAdam said the commercial launches would be pushed back at least six months to the second half of next year, with national coverage now scheduled to be completed in 2015 - two years later than originally stated.
McAdam chalked the delay up to a decision to go slow on LTE and "see what we need to do so we don't get ahead of ourselves in putting in capacity that we don't need."
But Verizon may also be taking a page from 3G's own chequered history, observes Caroline Gabriel, head of research at Rethink Research.
"This may not be the last revision of Verizon's timescales, and the story is a familiar one from the days of the European 3G bubble," Gabriel said in a research note. "Launching prematurely with ill-tested equipment and a shortage of devices would be worse than delaying roll-out."
And a device shortage is a real possibility. Qualcomm - which supplies most of the chipsets that end up in Verizon products - plans to have chips for LTE datacards generally available next year, by which time handset chips will only be available for sampling. With device product cycles typically lasting 18 months - and with Verizon's well-known policy of rigorous device testing to meet its high standards of quality - Verizon's LTE offering will have to get by on dongles or single-mode gadgets until the end of 2011, Gabriel says.
Device strategies
To be fair, Wimax has been dealing with the same issue. Wimax networks today ship mostly with either CPE for fixed-wireless offerings, or dongles, although a few laptops with embedded Wimax have also appeared in select markets. Dongles are a good entry-level approach to get consumers using the service, but Wimax proponents - as well as the GSMA, for that matter - believe that wireless broadband's success hinges on embedding it into more devices, starting with laptops and netbooks, but moving on to all kinds of consumer electronics devices like cameras, MP3 players and gaming consoles.
The problem, says Peter Cannistra, VP of market development at Clearwire, is that while CE manufacturers "buy the Wimax story", they're still hesitant to add to their BOM when they're already running on razor-thin margins and Wimax's business model is still a work in progress.
"Wimax players need to work harder to make Wimax compelling enough for the CE manufacturers to go forward," Cannistra says.
The need for multiple embedded Wimax devices isn't just about providing the variety of choice that drives 3G services today - it's also the key to competing against HSPA/EV-DO services now, and LTE in the future, according to Takashi Tanaka, president of Japan's UQ Communications.
In his Wimax Forum Asia keynote in April, Tanaka said that Japanese mobile users are generally unhappy with handsets as a mobile Internet device. "Most mobile users in Japan use mobile Internet [HSPA and EV-DO], but complain regularly about speed, the limited browser, limited content, small screen and small keyboard," he said. "Wimax can meet those needs, so why not deploy it?"
More to the point, however, users in Japan have lots of gadgets that they want to connect to the Internet, but mobile operators have difficulty with this because traditionally, a mobile service contract is tied to just one device. Consequently, UQ intends to offer multi-device mobile deals, where a flat-rate monthly fee allows a subscriber to connect, say, a laptop/netbook/dongle, an MP3 player and a gaming device.
The risks of all-IP
Such a strategy gets to the heart of the way in which Wimax could set itself apart from HSPA and LTE. At its core, Wimax is designed to be an open IP-based network. That means it has the freedom to try out new business models such as open access - i.e. as long as it's certified by the Wimax Forum, your device will work on our network, and you can include more than one under your service contract - and selling devices through retail consumer electronics stores. It also means the ability to offer more flexible tariff packages, where one can buy a day pass, for example.
To be sure, LTE is also designed to be all-IP and can exploit many of the same benefits. A number of vendors are already urging cellcos looking at LTE to make sure their core networks can handle the QoS functionalities on an end-to-end basis.
"There are several important aspects that are crucial to the LTE core," says Francesco Masetti-Placci, VP of solutions, strategy and marketing at Alcatel-Lucent China. "The first is latency requirements, less than 10ms, which cannot be supported with current architectures. Also, as you can expect a larger number of simultaneous users, you need more sophisticated processing power in the routers. And the architecture needs to be designed in a more distributed way to support the four classes of QoS under UMTS, which has to be mapped into the new architecture. Then policy based end to end QoS can be enforced."
However, for cellcos moving to all-IP, the potential snag may not be the technology to enable it, but the open-access business model that comes with it and runs counter to the existing cellco model of locked devices and two-year contracts.
To date, cellcos have been slow even to adopt IP models like flat rate access. UQ intends to exploit this to the limit, says Tanaka, who says that Japanese mobile users aren't just unhappy with the handsets as Internet devices - they also don't like being locked into contracts and see 3G operators as "the bad guys".
All of this represents a major and crucial opportunity for Wimax to differentiate itself now, says Gabriel of Rethink Research.
"Of course, LTE will be closer to Wimax in many key areas, such as all-IP and spectrum allocations, but with widespread LTE several years away, the priority for the Wimax community in 2009 is to put clear water between its own platform and HSPA/EV-DO, and demonstrate that the systems have very different functions, all needed in the mobile broadband world," she says.
4G: where's the money?
As telecoms continue 4G equipment rollouts, one question looms: Will wireless data revenues keep pace with services offered?
Many analysts say no, unless telecoms find a killer, money-making 4G service.
"I've always noted that what it really takes [to drive 4G revenue] is an application that needs to consume that bandwidth," said Mike Jude, a program manager of consumer communications services at Stratecast. "Currently, I think most LTE and 3.5 [later generation, faster 3G] to 4G deployment plans depend on this notion of mobile Web access or mobile data access as being the driver. But really, if you're looking at the total population of total users of wireless, not that many of them use that much data."
And not many are currently willing to pay much more just for higher bandwidth.
Instead, Jude and others believe, some basic models such as the all-you-can-eat flat-rate need to be fundamentally rethought or at least subsidized with high-margin services.
"Everybody's looking for services that they can provide [for] high-margin that would depend on that kind of bandwidth requirement," Jude said. "If it's a service that people want and it requires that kind of bandwidth, presumably they'd pay for that."
If wireless providers launch 4G services without a clear idea of exactly what those ARPU-driving services are, wireless providers may face the same crisis the wireline providers are facing: plumetting revenue-per-bit even as infrastructure upgrade costs skyrocket and entrenched consumers protest any change to the revenue model status quo.
H. Paris Burstyn, a senior analyst with Ovum, said wireless carriers need to embrace the death of the minute and move past it in terms of billing.
"Carriers are ... fixated by minutes of use, and that's a terrible way to think about digital service because it's not about minutes of use, it's about megabits, and a growing quantity of data," Burstyn said. "The key is to figure out how to move the payment metric from minutes to megabits, and then tiers of service beyond that, which spills into net neutrality, and dealing with power users."
Jude said that there is some time to adjust, particularly with 4G. So far, however, a rational business plan has yet to emerge that will match the costs of 4G capital and operational expenses with revenue, Jude said.
Whatever the answer, it won't be found overnight.
"It's a very complex problem, and monetizing it is the big issue," Burstyn said. "Moving the consumer and the billing paradigm to something that's linked to the amount of data you use is going to be a very complicated and evolutionary step-by-step kind of thing."
- Michael Morisy / SearchTelecom