With fourth generation networks coming soon, mobile operators must chose which standard they adopt.
From the euphoria and irrational exuberance that surrounded 3G when the first government licences were awarded at the end of the 1990s, high-speed cellular service deployment has progressed slowly. Some operators, such as O2 in the UK, have only just met their obligations to government to roll out 3G services. After receiving a slap on the wrist from regulator Ofcom in February, O2 managed in May to extend its network far enough to cover 80 per cent of the UK population.
Yet, even as the first 3G networks started handling user calls and data earlier this decade, work started on the next-generation infrastructure. The first 4G networks are likely to get underway soon, with operators now choosing the standard with which they will go.
"The interesting thing is that the technologies are, in a broadbrush way, quite similar," says Rupert Baines, vice president of marketing at wireless-chip developer Picochip. "The differences between them are tiny compared to everything that has gone before."
One thing became very clear early on. The 4G networks would use some form of orthogonal frequency-division multiplexing (OFDM) - the same basic technology as that used in Wi-Fi wireless networks and ADSL for wired access to the Internet. But that is where the agreement stops. As with 3G, vendors and operators are not aligning behind a single standard but are choosing to slug it out with similar but competing protocols. The situation is a bit simpler than it was for 3G, where the world ended up being faced with not one 3G protocol but three of them, largely split along geographic lines, although they were all based around the core technology of code-division multiple access (CDMA).
CDMA2000, based largely on technology from Qualcomm, became the preferred standard for the US and Korea. In Europe and Japan, wideband-CDMA prevailed. China has spent the last five years gearing up for its variant, TDS-CDMA, with networks expected to be serving users by the time the Beijing Olympics gets underway.
This time around, there are again three candidates for the wireless operators to choose from. However, observers do not give Qualcomm's successor to CDMA2000 much chance of success. Operators who were expected to stick with Qualcomm because they have CDMA2000 networks today have instead opted to go with the successor to W-CDMA or the wild card of WiMax.
Qualcomm has, to date, not announced any deal with an operator to use its Ultra-Mobile Broadband (UMB). US operator Verizon, which supports CDMA2000 today opted to use the Long-Term Evolution (LTE) of W-CDMA, as did Vodafone in the US. Sprint, the third-placed US operator has broken with its cellular tradition and decided to go with WiMax.
"It would be foolhardy to say that UMB is dead, but it's not looking rosy," argues Jason Chapman, managing vice president at Gartner.
The two likely remaining standards look similar on paper but there are key differences, says Chapman: "The back - ground to the technologies is from the IT side whereas LTE has very much a cellular heritage. They have different philosophies but are moving to a similar goal."
One difference between LTE and WiMax lies in timing. "WiMax probably has a threeyear head-start over LTE," says Baines. "LTE is coming in a couple of years and with a couple of years' more smarts. LTE has a very clever uplink, for example. But then this is always a game of leapfrog.
"The [WiMax] 802.16e standard was frozen in 2005. And the LTE standard is more or less there now: it is supposed to be finished this year. As it was frozen in 2005, it is pretty much real this year," Baines adds.
The later start for LTE means that the first networks using that standard will probably not get going until around 2010, with trials expected in the second half of next year, according Francis Sideco, analyst at iSuppli. "We will see testing by NTT-Docomo and Verizon in late 2010," he claims.
Gordon Aspin, chief technology officer of Camitri Technologies, says: "I doubt there will be a compelling LTE service in 2010. It will be another couple of years before we see anything compelling out of those markets. At the moment, the European presence seems a little weak: I don't see them rushing to implement LTE. But that is because they are looking to use their existing infrastructure."
The 3G slog
The problem for many of the operators with existing 3G infrastructure is that it took a lot longer than expected to get customers onto the new system. A timetable for 4G that sees it get underway in the next few years looks unnecessarily rapid when you consider that, for a number of operators, they have not finished with 3G yet. In China, they have barely started.
"Before we get to 4G we are likely to see an uptick in 3G basestation deployment," says Julian Hildersley, vice president of strategy and marketing at poweramplifier specialist Nujira.
Chapman says: "The thing that people need to remember is what is driving the shift. When we were at 3G, the philosophy was: build it and they will come. Here's a fast network: you can do video on it. But people didn't want peer-to-peer video.
"It is a very different migration path that the operators are taking now: the driver is wireless broadband," Chapman adds, pointing to the rollout now of the data enhancements to W-CDMA: the High-Speed Packet Access (HSPA) modes for fast download speeds (HSDPA) and then faster uploads (HSUPA).
"Operators do not want to be rolling out a new infrastructure just because that is the next thing. HSDPA and HSUPA are providing the technology that operators and consumers demand," says Chapman.
Baines agrees: "For most people who have got HSPA networks there is a few years of life in them. They don't have the compelling drive to move."
LTE will provide higher datarates and greater efficiency in the long term, particularly when it comes to terminal battery life. Baines points to the novel uplink that LTE uses as an example. "Terminal battery life will be better at the expense of some complexity," says Baines, adding that latency is improved on LTE compared with 3G, which will be useful for games as well as calls that make use of Voice over Internet Protocol (VoIP).
However, without WiMax to chivvy them along, observers believe that we would not be looking at LTE networks getting underway before the end of the current decade.
"Left to its own devices, we probably would not have seen LTE until 2012 or 2013," Sideco claims. "WiMax has pushed LTE to accelerate past what it would have done organically."
Would-be LTE providers see WiMax as competition because of the people embracing it. Without LTE in place, the slightly more mature WiMax standard looks much more attractive to some, says Baines. "As a greenfield operator, you will be more interested in a technology that you can have today.
"The combination of what investments you have got and how quickly you need to deploy largely make the direction very straightforward. A lot of CDMA2000 EV-DO people will go for WiMax. That explains Sprint, KDDI and the Koreans," Baines adds.
Sideco explains: "In general, WiMax is embraced by those who want to change their lot in life. Sprint has been at number three in the US for a while. WiMax probably has a one- to two-year head start, so they are trying to differentiate earlier on. Some of the other guys are trying to get the return on investment."
Sideco says he sees WiMax being important in markets where 3G does not have a hold at the moment, especially those with large sparsely populated rural regions. Countries such as Russia, India, Pakistan, the African nations and Australia look to be good candidates for WiMax. So, although there is not the obvious Atlantic divide that split 3G apart this time around, geography continues to play a part in the evolution of wireless communication.
In a good many territories, WiMax and LTE coverage will overlap. For example, Intel bought a licence that will allow the number-one chipmaker to build a network to provide WiMax services. The company has an interest in driving WiMax as it aims to expand the position it has built up using Wi-Fi onto the newer standard. The Intel move will see it compete with operators such as Telia, which is likely to move to LTE. However, operators deploying the different systems may not compete directly.
"There has been a transition in how WiMax proponents are positioning themselves," says Sideco. "Initially they said: 'Let's take LTE head-on'. Now they are saying it is complementary."
The issue for WiMax is how to handle handovers as a user moves out of range of one basestation and into contact with another. With years of experience in the previous generations of cellular telephony, LTE has something of a head start in this area. "Handover and those things are a real pain in cellular," says Chapman.
Hildersley agrees: "The greenfield operators without a lot of experience think it will be like working with 802.11 [Wi-Fi] hotspots. The reality of providing wide-area coverage with complex handoff is harder than they expect. A lot of the complexity in protocols such as 3G is there for a good reason. You have to deal with that sooner or later."
The WiMax standard is far less defined in how the core network operates, says Hildersley, which affects situations such as handovers. It may lead to lost connections for users moving at speed. In practice, this may not be a problem for WiMax operators who are likely to focus on nomadic rather than mobile use, much like the situation with Wi-Fi hotspots today.
"If you look at those use-cases, 80 to 90 per cent of the usage is nomadic," says Sideco. Most of the time, the operators are not expecting users to be travelling at 100km per hour in a car or train. "It has its mobility aspects. But when you think about the use-cases, the mobility is probably not going to be the same as with LTE."
The migration to 4G is not going to be a rerun of 3G with countries picking one standard over another. However, the situation for users may be more confusing than it is now as the two main 4G standards will be duking it out in the same territories with only the lessdeveloped nations perhaps seeing WiMax services on their own. There is talk of trying to converge the two to improve interoperability, according to Aspin. But until that convergence comes, you can expect to see a lot more conflict between LTE and WiMax.
In their respective standards groups, LTE and WiMax are each acquiring techniques favoured by the other as operators try to work out how they will be able to deploy 4G services. The problem the operators face is that radio spectrum is highly fragmented and the situation is not going to get better any time soon.
Julian Hildersley, vice president of strategy and marketing at Nujira, says: "One of the issues across the industry is that of frequency bands assigned to the specification. There are 14 bands proposed today for LTE and others are coming up fairly regularly. I see quite a lot of regulatory authorities want to reallocate UHF spectrum. So, there is a risk of regional fragmentation both for handsets and infrastructure."
One potential issue with LTE for operators is that the current version of the standard calls for frequency-division duplexing (FDD). However, in some cases, it may be difficult to obtain the matched spectrum that would allow uplink and downlink channels to use different frequencies. So, the LTE standard is likely to wind up with a timedivision duplexing (TDD) option to provide operators with more flexibility.
Conversely, WiMax operators are pushing for an FDD option to use in place of the TDD version to allow them to use the protocol on frequency bands that require FDD to be used. One example is the paired-spectrum 700MHz band that the US decided to auction.