(CDMA2000) A Winning Hand?
Vaughan O'Grady
Evolution Issue 4
Having survived a few early reversals, it looks like CDMA2000 now holds all the cards: a 3G system with millions of subscribers, apparently impressive rates for high-speed data and, most importantly, greater voice capacity than ever before. But where does the technology go from here? Will its next phase find an audience? And what happens when WCDMA finally rolls out on a larger scale?
It may have seemed a disappointment at first when the cdmaOne community failed to buy into wideband CDMA2000 3X, but the result seems to have been more a boon than a compromise solution. The 1X path has now been defined as the primary evolutionary path from cdmaOne, and been granted 3G status by the ITU. More importantly, it has gained acceptance in the marketplace, stealing a march on its wideband rivals with roll-outs planned or under way in numerous territories - but notably the two leading ones in terms of subscriber numbers: Korea and the US.
However, the motivation for going ahead with 1X may differ from one operator to another. For many the important point is that they are gaining, in the same bandwidth, "a significantly greater capacity for voice on CDMA2000", as Dr Irwin Jacobs, chairman and CEO of Qualcomm, puts it. Phillip Redman, research director in Gartner Research, explains: "CDMA2000 is much more aggressive in controlling capacity on the network... once there's high penetration of CDMA2000 devices on the network that are 1X-compatible, it could be anywhere between 1.2 to two times more capacity on that network for voice because it's that much more efficient." This, as Mike Iandolo, VP CDMA Product Management of Lucent, puts it, "from our customers' point of view, by itself justifies the investment they are making".
Also, 1X offers high-speed data, which leaves operators with an interesting conundrum. Should they stick with 1X or, now that they've made the major outlay, do they think demand will justify spending a little more buying into the next stage, EV-DO, available now? Iandolo is careful how he characterises this choice: "The sequence that the US carriers are going through is that the first focus this fiscal year is to roll out 3G1X ubiquitously and to set the stage for the voice capacity improvements and introduction of data services," he suggests. Which carriers choose to go for 1xEV-DO, then, depends largely on whether voice enhancement is their sole interest at present and whether they believe high-speed data on 1X is all they need right now.
Christine Loredo, senior analyst, Mobile Wireless Research, of the Strategis Group, notes that US carrier Verizon, which launched a limited 1X service earlier this year, is already testing EV-DO in a few markets and suggests: "They're ready to go as soon as possible - the reality is early 2003, but if they can get it out by the end of the year, they will." If they don't, she says, US-based EV-DO may this year come from an unexpected source - a small operator called Monet Networks that is offering CDMA2000 "as a DSL substitute in second and third tier markets that don't have any type of broadband service" - proof that there is already some lateral thinking going on around CDMA2000 evolution.
As for Verizon's biggest competitor, the Sprint PCS plan is to launch Release 0 of the standard nationwide in the third quarter and then move onto EV-DV at an unspecified future date. Oliver Valente, CTO at Sprint PCS, says that this is because "DV will allow voice and data on the same spectrum, which really offers operators more flexibility". By contrast, "you would have to dedicate your spectrum for 1xEV-DO, and that spectrum would only support data". The benefits, of course, are that it would be able to peak at speeds of over 2Mbps, but Sprint feels, says Valente, that "there is not an application out there demanding that kind of speed at the moment".
Advanced data services Korean operators may, of course, disagree. SK rolled out EV-DO early this year and KTF has been hard on its heels. Why? Richard Ferguson a co-head of regional telecoms research for Nomura Asia Pacific, argues: "It's a combination of demand plus government encouragement," pointing to an already high broadband penetration and continuing fascination with wireless and wireless services only really matched in Japan, both of which encourage a belief in the viability of a move to the sort of advanced data services EV-DO could offer.
But Korean operators hold licences to roll out WCDMA and CDMA2000 3X.
Will they move beyond EV-DO? It's conceivable, says Ferguson, that if EV-DO does the job required of it, they might not see it as viable to spend money on a new system to which they may have to migrate otherwise satisfied customers. Ferguson suggests that they will say: "We don't invest any further until we see if this model is likely to give us the returns we want from it." What happens to the spectrum in that case is a matter for some speculation.
Whatever the future brings, present evidence of 1X performance is encouraging.
It's already available from over a dozen operators in Asia and North and South America and Hank Menkes, vice president, Global Spread Spectrum Systems Engineering, Lucent Technologies, says that, along with the improved voice capacity, "from the data standpoint, in a loaded system with multiple users on it, we're seeing data rates that are comparable to some of the best dial-up connections that you can experience here in the US... data rates in the range of 40 to 60kbps on average with peak data rates that go as high as 153kbps." EV-DO can improve on that, of course. And, depending on who you talk to, EV-DV matches DO's data speeds or exceeds them, but in the same channel - hence Sprint's interest.
However, if other operators care to follow Sprint's example, they will be delighted to note that the standardisation issues that bedevilled CDMA2000 EV-DV are close to resolution. As Ed Tiedemann, senior VP for engineering of Qualcomm and 3GPP2 chair of the Physical Layer Group, accepts, the process has been a touch tortuous: "There were originally eight proposals.
Those consolidated into two real groups: one was the 1Xtreme group [led by Motorola and Nokia]. The other was a group called L3NQS, which was Lucent, LG, LSI, Nortel, Qualcomm and Samsung. The two groups had sizeable teams working on analysis, simulation, study and so on as part of that.
In October [2001] there was pretty much an agreement to go with the L3NQS proposal with some modifications, as often happens in standards processes, to bring in some concepts from the 1Xtreme proposal." Essentially what that means is that a substantial standards definition of 1xEV-DV, which is revision C (of CDMA2000), has now been put together. It's now gone out for Standards Development Organisation review and comment, which allows the main organisational partners - TIA (North America), TTA (Korea), CWTS (China), ARIB (Japan) and TTC (Japan) - to consult their membership.
Comments have come in, technical issues will then be resolved and the first release of the standard will be published but, however eagerly it is awaited in some quarters, don't expect commercial products soon. According to Iandolo, the expectation is, "that you'll start seeing commercial products for EV in the 04/05 timeframe."
So, by the time EV-DV comes to market, WCDMA should have found an audience.
Few save the most hardened CDMA2000 supporters deny that WCDMA will probably win out in terms of subscriber numbers given the strength of its GSM operator base. Loredo certainly thinks so but notes: "I think it's going to happen a lot later than we originally thought because of the technology problems - we don't even have handsets yet. It'll be dominant but it's not going to happen in 2004 - and probably not 2005." Redman adds that, even though there is "no clean upgrade path to GSM from EDGE to WCDMA - it's all new equipment at each step", political and regional issues will probably keep GSM operators on that path.
And EDGE? Again 1X has the advantage of market performance, but Dr JT Bergqvist, senior vice president and general manager, IP mobility networks, with Nokia, is unconcerned: "It will outperform CDMA2000 in terms of data rates available to customers," he says, adding: "We can measure that in the market by the first quarter of next year." Lucent's Menkes is not so sure: "If you normalise all of this so that you're talking about using 5MHz of spectrum - you can get three times the data throughput with CDMA2000 EV-DO that you can with EDGE in the same 5MHz of spectrum. So in terms of capacity or throughput we're not even talking apples and oranges they're so far apart." Dr Jacobs at Qualcomm is also less than convinced: "It has been published that because you need a higher signal to noise ratio the maximum coverage of a base station is less for EDGE than it is for GSM or GPRS. But then people say: 'Most places I've had to build more base stations because I need the capacity'. But what they don't recognise is they're still limited by the interference that one base station causes in an adjacent base station - you can't simply arbitrarily increase your signal power."
He also believes that 3X's day has gone. "I don't think anybody's going to build the equipment or a chip for that at this point because it turned out that with 1X and now 1xEV-DO that there's enough spectral efficiency to do everything in the existing bandwidth." And if you won the extra bandwidth? "As far as efficiency and data rate in so much per megahertz, 1X does the job; you just put in as many carriers as you need at any time."
And the foreseeable future? Dr Jacobs cites the benefits of the multiple handset antenna technology that he believes 1X allows combined with the next version (Selectable Mode Vocoder) of the CDMA2000 vocoder. "Between those, we'll get about a factor, above 1X, of two on voice and some improvement in data," he says. Menkes cites capacity as the main issue and thus "you're going to see things like intelligent antennas start to become part of each of the vendor's base stations which would then allow us to add even more subscribers on this system."
The next revision On the standards side, Ed Tiedemann suggests that once whatever bugs are apparent have been ironed out of Revision C of 1xEV-DV, it will be a matter of moving into what's going to be in the next revision, Revision D - whatever that may be. "Nothing concrete has been locked down but there's been discussion of providing some enhancements on the reverse link," he hints.
Other concerns which will complicate matters for both 3GPPs are waiting in the wings. It's a sign of the times that the SIP (Session Initiation Protocol) Forum and the IETF (Internet Engineering Task Force) are now influences on where mobile standards are going because of plans for the mobile internet, and hopes one day of, as Tiedemann puts it, "the general agreement that we should be moving towards one network - particularly in the direction of having a common IP-based network." Although such a transition will, he believes, be slow, there is a need to adapt established IETF standards and protocols for wireless. SIP especially is rather demanding in terms of bits, so compression is also an issue that needs to be tackled, explains Tiedemann.
Meanwhile 1X is being rolled out or planned, not just in the US and Korea but in such diverse territories as Canada, Brazil, Australasia, Romania and China. The hope of its proponents for the longer term is that data will take off, and eventually the move will take place from circuit to all-IP, bringing even more enhanced applications.
That is pretty much the vision for UMTS too, Menkes notes, but he adds: "I think with the fact that 3G 1X is now in the field and data's starting to take off we'll probably get some experience with these features a little sooner than we would with UMTS at this stage." |
