Ever since the International Telecommunication Union created its IMT-2000 family of 3G technologies ten years ago, 3G has come in two basic forms: FDD and TDD. FDD offered separate spectrum bands for uplink and downlink traffic, while TDD offered upstream and downstream traffic on the same band.
But while FDD-based technologies like W-CDMA and CDMA2000 dominated 3G rollouts (the former more so than the latter), TDD-based TD-CDMA remained generally unloved except by vendor IPWireless and its small customer base, while the other TDD option, TD-SCDMA, was China's homegrown technology that most industry players regarded as a China-only technology with no realistic chance of ever seeing deployment outside of China, although Taiwanese operators FarEasTone and Vibo Telecom began small-scale trials of TD-SCDMA late last year.
However, TD-SCDMA - which went fully commercial in China last year and is now available in 238 cities from China Mobile - is suddenly looking primed to go international in the form of its next-gen version, TD-LTE, which has the full backing of the 3GPP as part of its overall LTE standards effort. That doesn't automatically guarantee greater operator interest in TDD outside of China - where, like its 3G predecessor, TD-LTE is guaranteed a home. But the past couple of months have seen a flurry of announcements around Asia regarding TD-LTE's viability.
Not unexpectedly, China Mobile signed a deal in April with FarEasTone to jointly develop a TD-LTE trial in Taipei. Much more surprising was Qualcomm's announcement last month that it would enter the long-awaited spectrum auctions in India to bid for broadband wireless access (BWA) spectrum in the unpaired 2.3-GHz band, with the intention of using it for TD-LTE, rather than Wimax. Later that month, Japan's Softbank Mobile - which gained access to 2.5GHz TDD spectrum in March after acquiring a stake in failing PHS operator Willcom - said TD-LTE was definitely on the table as an option. And perhaps tellingly, as TDD vets like Huawei, ZTE, Motorola and Nokia Siemens Networks have been issuing progress reports on TD-LTE developments, Ericsson announced a tie-up with Datang Telecom to integrate Datang's TD-SCDMA RAN gear into its 3G portfolio, ostensibly to beef up its China business, but also to develop TD-LTE solutions.
Such is the sudden excitement over TD-LTE that it's not only expected to see rapid take-up once commercial deployments starting next year, but will also comprise a sizable chunk of the entire LTE market by 2014, says Mike Wang, NSN's general manager for Taiwan, Hong Kong and Macau.
"It took W-CDMA six years to reach 100 million subscribers, but LTE will hit that in four years, and we project that TD-LTE will account for 40% of LTE subscribers by that time," he told Wireless Asia.
What's going on here? What's TD-LTE got that its TDD ancestors didn't? Part of the answer lies in spectrum availability and changing market dynamics. However, a major key to the puzzle is, ironically, LTE's arch rival Wimax, which is not only helping make the business case for TDD in general, but - even more ironically - could serve as the platform for future TD-LTE deployments.
TDD then and now
To understand TDD's appeal in 2010, it's worth looking at why TDD didn't catch on during the 00s despite being commercially available. Part of the reason the industry favored FDD was to do with technology, explains Dr Ray Owen, head of pre-sales for the Mobile Devices and Home division of Motorola Asia and general director of Motorola Vietnam.
"FDD separates the uplink and downlink, which is much simpler from an engineering standpoint than TDD where both links are combined," he says. "That also helps in terms of battery life. Also, W-CDMA was really designed as a voice network that supported data and video, and voice and video perform better with symmetrical links. TDD is great for pure-play data, but 3G operators weren't focused on that model then."
Another issue, he adds, was that with everyone focused on FDD, TDD was only likely to be deployed if dual-mode FDD/TDD devices were available. "It's not impossible to put FDD and TDD in the same handset, but it is very difficult from an engineering point of view," says Owen.
By contrast, today chipset makers like Qualcomm and ST-Ericsson have FDD/TDD chipsets on their roadmaps, with China Mobile pushing for a dual-mode strategy as part of its TD-LTE plans.
Related to that is the close development of both FDD and TDD-based LTE on the network side, says Isaac Liang, international TD system marketing director for ZTE.
"FDD and TDD [LTE] are like two brothers, like twins - they are very similar in terms of coverage, throughput, spectral efficiency, and so on," Liang says. "Many vendors using the same platform for their FDD and TDD LTE products, and chipset vendors in their roadmap will have dual-mode FDD/TDD wireless chipsets. These are incentives for operators."
Another factor is the rapid rise of dongles and IP-savvy smartphones like the iPhone, which have driven mobile data traffic to such levels that operators in many markets are hungry for new spectrum - and they aren't necessarily going to wait for paired LTE spectrum to be freed up, says Tom Gruba, Motorola's senior director of wireless product marketing.
"Between iPhones, netbooks and flat-rate pricing, operators are finding they don't have enough spectrum, which is one reason they want 2.6-GHz LTE, but they'll also use all available spectrum they can get," he says. "They're seeing all this TDD spectrum being used in China, the US and in southeast Asia with 2.3 GHz. So there's all this opportunity for spectrum, and China has broken that inertia to get things moving."
Julien Grivolas, principal analyst for networks and technology at Ovum, agrees, noting the example of Softbank acquiring Willcom primarily for its spectrum. "They're not sure what they'll do with it, but what they want is the spectrum. You will not say 'I'll only take this kind of spectrum', you'll take what you can."
Wimax now ...
The other key difference, says ZTE's Liang, is that TDD has a major champion that it didn't have before: Wimax.
"The Wimax Forum has been pushing their technology, which has led to releasing 2.3 GHz unpaired spectrum in many markets, many of them markets where LTE spectrum isn't yet available and may not be for some time," Liang says.
Which sounds like better news for Wimax than for TD-LTE. Both fixed Wimax (802.16d) and mobile Wimax (802.16e) have a considerable head start on TD-LTE in terms of commercial deployments and a fairly robust ecosystem. And while the GSM Association has repeatedly argued that cellcos need to focus on the LTE roadmap and not get distracted by "interim" niche technologies like Wimax, the Wimax Forum and its members have maintained that there's no reason to wait for LTE in any form when the demand for IP-based wireless broadband is there now.
"When Yota says they spent $12 million to get 350,000 subs, that's a customer acquisition cost of around $34," says Gruba of Motorola. "When you look at what it costs to acquire cellular subs, which is several hundred dollars, the message is 'win share now', because when a second competitor comes into that market, regardless of the technology they use, their customer acquisition strategy has to be much different. You've either got to offer more or be price-conscious."
Ashish Sharma, marketing VP for Wimax vendor Alvarion, adds that TD-LTE is a lot farther off than it looks.
"TD-LTE doesn't have the ecosystem in place, and you need at least four or five years for that to develop, which Wimax has already done," he says. "Also, the first market that will deploy it on any decent scale will be China, which is a different spectrum band [2.5 GHz] than other markets and has to be backwards-compatible with TD-SCDMA, which will fragment ecosystem developments."
... TD-LTE later?
Interestingly, however, the "Wimax now" argument has morphed into "Wimax now, TD-LTE later" as vendors who sell both Wimax and LTE - such as ZTE, Huawei, Motorola and NSN, among others - point out that Wimax and TD-LTE are currently not all that different in terms of hardware.
"There's about 80% to 90% technical overlap between Wimax and LTE," says Mohab Khattab, group VP of Business Development/Legal for wi-tribe. "Both are based on OFDM, both are all-IP, they can use the 2.3/2.5 GHz bands, they employ advanced antenna designs and have similar speed capabilities."
Wang of NSN makes the same point. "We want to protect our customer investment. In the BTS side, many components that can be used for Wimax and LTE, so it's an upgrade path. The whole TD-LTE ecosystem will be ready by 2011 or 2012, so you customers can roll out Wimax now and have an upgrade path to TD-LTE once that ecosystem is ready."
Put another way, today's Wimax networks are - potentially - tomorrow's TD-LTE networks, which in turn strengthens the case for deploying Wimax today, says Richard Ye, brand director of product R+D system and wireless product operation at ZTE.
"If operators see a solution that can run both Wimax and TD-LTE, for sure they will launch Wimax first, because they can make money now, and use a small investment to reconstruct for TD-LTE," Ye says. "You get to keep the subscribers you built up with 16e, and then if you upgrade to 16m or TD-LTE, it's a software decision. You can even run both, if you want to do so and have enough spectrum."
Khattab of wi-tribe adds that the amount of extra investment needed to make such a transition isn't all that high. "When you look at all the costs involved in building a Wimax network - the core, the backhaul, site rental, human resources, utility costs, the whole thing - the actual Wimax part is just the base stations and the CPE, which only accounts for around 12% of your investment," he says. "That's all you really have to change to migrate to LTE, so from an investment point of view, that's relatively painless."
Not so simple
To be sure, Wimax-to-TD-LTE migration is not endorsed by either the GSM Association or the Wimax Forum, both of which are trying to build scalable ecosystems and massive subscriber bases for their own technology. But, says Wang of NSN, "When you talk to individual operators, they are highly interested in TD-LTE, because it gives them a mainstream evolution path. And if you ask Wimax operators in particular where they want to go two or three years down the road, they all say ultimately LTE."
On the other hand, GSMA and Wimax Forum officials aren't the only ones skeptical of the viability of a Wimax/TD-LTE evolution plan.
For a start, says Fabricio Martinez, group services product director at network planning consultants Aircom International, the 12% investment figure from wi-tribe is far too low.
"Even in the most optimistic of scenarios we see around 18% to 20% as the lowest possible loss, and with detrimental effect on the quality of the network being offered," Martinez says.
"In order to achieve even an 18% saving, the network would have to reuse the core network, backhaul, Tx and full site and antenna system," he adds. "Crucially, this network would offer no roaming capability and no voice fallback. All upgrades would have to be software-based and the implications would be significant."
Ovum's Grivolas adds that even if a given Wimax operator has the right equipment and spectrum amount to support a smooth RAN software upgrade - which won't always be the case - that software upgrade doesn't apply to devices.
"All of the devices will not be able to evolve their software, so there will be a huge spend there to subsidize new devices," he says.
In any case, he adds, with the Wimax Forum continuing to champion Wimax, as well as vendors like Alvarion who are focused primarily on Wimax, Wimax as a technology isn't going anywhere. "I don't believe that in 2015 there will be no Wimax networks."
Phillip Solis, practice director at ABI Research, notes that while vendors are indeed working to enable Wimax/TD-LTE migration, as well as dual-mode Wimax/TD-LTE chipsets, the migration message is more for the benefit of investors than operators.
"Investors are seeing TD-LTE making lots of headlines, which makes them look at Wimax operators and ask, 'Why should I invest in you when you're running Wimax?' So operators can say, don't worry, we're covered, we can make the switch if we need to. It's like insurance, which is actually good for Wimax because operators and investors are more likely to take a chance on it if they have that switchover option later."
That said, Solis adds, the Wimax and LTE landscapes will be complicated by fragmentation between FDD, TDD and related spectrum bands, which will make dual-mode ambitions difficult, and that will work against TD-LTE.
"People are assuming a lot of market synergies around TD-LTE that actually don't exist," he says. "It's complicated, but you've got LTE development going on for FDD and TDD, then there's all this band fragmentation with 4G, because you've got 2.6 GHz, refarming in the 3G bands plus backward compatibility, and digital dividend spectrum coming up," he says. "And in some markets like the US, you'll need multi-mode support for EV-DO in all that, and in China, TD-LTE chips will be TD-LTE/TD-SCDMA chips. Software-defined radios help and bigger RF front ends help, but you'll still see a lot of fragmentation that applies to TD-LTE."
RELATED ARTICLES ON TD-LTE