LTE might have won the war against all other competing technologies on the road to 4G, but the platform still faces several major challenges before it can actually present itself as a truly next generation technology.

For starters, while LTE networks are being trials across the world and commercial equipment is being rolled out by several forerunners – NTT DoCoMo, Verizon, TeliaSonera, to name a few – it remains very much a data-only networking platform for now.

According to GSMA director of technology, Dr Dan Warren, the actual implementation of voice on LTE networks has yet to be finalized.

“The interesting discussion in this particular space is around voice on LTE and how that’s going to be deployed,” Warren said, pointing to a recent formation of a group called the One Voice Initiative – made up of Vodafone, Verizon, AT&T, Telefonica and Telia Sonera and it is backed by Ericsson, NSN, Alcatel-Lucent, Nokia handsets, Sony Ericsson and Samsung – to work on the problem.

“What they’ve done is taken all the 3GPP work on IMS and MMtel and profiled it for LTE. What they are trying to do is say, what needs to be done is take a common implementation on voice for voice on LTE to work,” he said.

“At the same time, what we are doing at the GSMA – we’ve had work going on for about 3 months, which – they are coming from the bottom up and taking the specs and profiled them, we are coming from top down – so saying that GSM works on the basis of a massive ecosystem, everybody uses the same interface, so I can take any device and go internationally. So to make roaming work, to make scale work, you need a single common implementation for things like voice.”

He cautioned: “When you get to voice over LTE, there’s a danger that that is not going to take place, unless someone like GSMA stands up and says, that’s what we’re going to do. The One Voice Intiative I think is a good bottom up approach and we’re going to come top down, and I think we’ll meet somewhere in the middle.”

Right now, he is working on a problem statement and project outline for the GSMA, but notes that any results probably won’t arrive for some time.

“This is going to be a significant piece of work and it will take a long period of time. I can define the profile in five days. It will take six months to get people to agree that it is right and it will take another 18 months to get the industry to back it. The technical work is actually not that complicated. It is a matter of getting everyone to line up behind it,” he explained.

WHY VOICE ON LTE: While there seems to be no immediate need for voice over LTE, as many operators still see the technology simply as a platform for off-loading data traffic on their increasingly congested HSPA networks, Warren says that voice will be necessary in the future as more real time applications are deployed on networks.

“If you don’t give people voice on LTE, on handsets, than you don’t have the problem to worry about,” he points out. However, that means limiting the application of the LTE to pure data transmissions.

Today, without support for voice on LTE, users will have to switch back to 2G or 3G to make voice calls. More importantly, they have to terminate their LTE session in order to do that.

“The drive behind it (voice over LTE), is that if you are going to have LTE as a technology and particularly LTE handsets, the mechanism to initiate voice calls, when you have an LTE bearer set up – it’s something called CS Fallback.”

“The way CS Fallback works is you initiate signalling within the LTE path if you have an active LTE radio session, and the LTE session has to be terminated and you cross the device back onto 3G or 2G to take the voice call, which is not great if your LTE session is doing something that’s an interactive media session of some sort,” he said.

“On the basis that you might want to take something that might be a sharing applications, maybe a laptop based application, it might be video call, it might be something where you want to add a voice call on top – it that means that everything you’re doing has to be terminated to make the voice call work, that’s not a great user experience.”

In this way, he adds, voice on LTE is not so much about voice as a standalone service, but rather “what voice adds to everything else that you can use the LTE bandwidth for.”

SPATIAL DIVERSITY: Even if the industry works out how to deploy voice on LTE, another major obstacles stands in the way of it becoming a full feature service like 3G and 2G.

The problem, Warren points out, is that MIMO, which is used by LTE to boost data rates, requires spatial diversity – basically a gap between two antennas – in order to avoid interference, thus making it hard to keep the size of handsets down.

“And the lower the frequency, the more spatially diversed the antenna has to be in the device. So if you are Verizon, you’ve got 700MHz spectrum, that’s quite a big gap (needed to support MIMO on the handset) – and it may be a gap that is bigger than a handheld device. They are going to have to put something in the middle to make sure the two antennas don’t interfere with each other, but that’s going to be some time away,” he said. “The form factor is going to be quite complicated.”

Even if voice does get implemented on LTE, how it drops back down to 3G and 2G networks is another challenge.

“As you fall out of LTE coverage, you don’t want the voice calls to terminate, so you have to implement something call Single Radio VCC (voice call continuity), which is very technically challenging and is taking a long time to define,” Warren points out. “There’s a lot of challenges still around voice on LTE and voice interaction between LTE and existing 3G and 2G networks.”

At the end of the day, he concludes: “If you don’t give people voice on LTE, on handsets, than you don’t have the problem to worry about.”

BACKHAUL CHALLENGE: Even if the industry gets voice right, there’s one more challenge left for LTE – backhaul.

According to Warren, the spectrum efficiency gains of LTE will enable hundreds of broadband users from a single cell site, hence escalating dramatically the need for backhaul.

“The applications that were enabled by mobile broadband changed the traffic pattern of the end users, so it fundamentally means that people can do video streaming instead of just web traffic, so more people active, for longer, with more bandwidth, so it’s a triple exponential problem for backhaul,” Warren explained.

“High data rate copper, preferably fibre, is fundamental for this stuff. There’s a lot of throughput and a lot of users that the network needs to support.”

One possible solution is the deployment of femto cells, which would effectively migrate a lot of the indoor traffic onto the broadband network. But even here Warren warns, the situation is far from resolved.

“One of my concerns on femtos is clock synchronisation, radio synchronisation, which requires all cells to be transmitting in a synchronised fashion. I don’t believe DSL lines allow femto cells to synchronise in the same way, which potentially causes interference, white out zones,” he said, adding that multi-operator households – where different members of the same family are on different operators , as well as MVNOs and managed networks – where the the operator doesn’t directly managed the networks themselves – will add further complexity to the scenario.

Tony Chan