The future of 4G wireless, according to Ericsson

3G networks are now well-established, and the mobile operators are promoting them as a better alternative to Wi-Fi. Thoughts are turning to 4G, the next generation of wireless broadband, and the promise of 100Mbps and up.

The battle for 4G is between two technologies, WiMax — developed by Intel — and LTE, the long-term evolution of 3G promoted by the existing mobile industry. Both have high performance; WiMax was first out of the gate, while LTE has huge advantages in being the nominated heir of the current cellular industry — now at three billion customers and counting.

There are also questions about how and when any new standard can reach the market. Some possible spectrum is tied up with existing technology, and won't be available until analogue TV signals are switched off or older mobile-phone spectrum gets 'refarmed', and while the UK has an auction at 2.6GHz that's being delayed by legal appeals.

Some vendors believe LTE should be delivered through indoor base stations known as 'femtocells', reducing the cost of incremental rollout but seriously reducing the number of large outdoor base stations that operators will buy.

Ericsson is one of the biggest backers of LTE, and notably sceptical about both WiMax and femtocells. We spoke to John Cunliffe, Ericsson's chief technology officer for the UK.

Q: LTE is the main thing we should talk about. It's a fourth-generation phone technology. It's fast, it has Mimo [a multiple channel antenna technology] and OFDM [the modulation scheme]. We are now being told that it has fought off the challenge from mobile WiMax — unless you know something different?
A: I would concur. But there are places where WiMax is appropriate.

We don't build WiMax products — we decided not to because we think it is going to be a minority market. We install it, though; in the UK we are installing WiMax for Freedom4 (formerly Pipex). WiMax is simply a radio technology, so networks need backhaul and core network, plus equipment for authentication, authorisation and accounting, as well as the base stations.

However, LTE is coming from an ecosystem of over three billion devices. If you look at the volume and size of that mobile-broadband market, you'll get a perspective on the size of the WiMax market.

How fast will LTE be?
We've been doing some drive testing in Sweden with mobile clients. We were getting a peak of 154Mbps, a mean of 78Mbps and a minimum of 16Mbps download speeds.

So when is LTE going to happen?
There will be trials before the end of this year, and base stations ready for commercial service before the end of next . These will be multi-standard. The RBS 6000 we launched at Mobile World Congress this year will do 2G, 3G and LTE.

Shouldn't operators wait for LTE handsets that can use the network properly?
They will probably roll out in advance of the handsets, so when the handsets are available, they are ready to go. In any case, handsets will start to appear in 2009.

What will happen with 3G and HSPA, while we wait for LTE?
There is a roadmap for HSPA. We have 7.2Mbps in the UK at the moment, coverage is good and plenty of handsets do it. With high levels of modulation like 64QAM and Mimo, we can get 42Mbps out of HSPA… and even 80Mbps with further optimisation. It is now being promoted as an alternative to Wi-Fi. It is easier to set up, and has roaming. New dongles don't even need a CD — the software's built-in, in flash.

If there is still life in HSPA, will the economic downturn delay LTE? We've had a downturn in the telecoms industry before, related to the internet bubble. This time, banks are involved, so operators wanting to upgrade may find it hard to get the credit to buy a lot of new equipment.
Conceivably — but I probably shouldn't comment.

Do people really need faster networks for mobile data? We can all do email and Facebook very happily with what we have.
If you look at what drives speed, it is always video content, whether on mobile or fixed lines.

We have a consumer panel, in which we interview 36,000 people in 11 countries, who use fixed or mobile broadband. I looked at their video usage, and found that, if you look at people...

...who use video content several times a week, 36 percent will use video from YouTube, 26 percent use video clips, 16 percent share videos and 12 percent buy video.

That is on both fixed and mobile networks, but there is a move towards mobile. We know that in 2009, the number of people accessing broadband by mobile networks will be greater than the number on fixed networks. By 2013, 75 percent of people will use mobile broadband.

If they have to provide a lot of video downloads, will operators have vast problems upgrading their backhaul to meet the demands?
I'm not at all concerned about the technology. It's there to keep up. From the point of view of affordability, it may be more of a challenge. But BT has agreed to provide Vodafone with 60Mbps backhaul on its 21st century network, with an option to increase that to 80Mbps. That's a landmark — a step change in the amount of bandwidth available to serve mobile data.

The move to LTE is a move towards data and data is usually done indoors, where traditional 'macrocell' base stations give poor coverage. Will that be a problem?
This depends which part of the spectrum is used. Penetration into buildings is a problem at the higher end of current spectrum, say 2.6GHz. But there is also spectrum from the digital dividend and spectrum re-farming, which we think should be made available for mobile applications. We think that would be really appropriate for LTE. This could provide broadband connectivity in rural areas, to people who wouldn't have it otherwise.

Speaking of 2.6GHz, every time I look, Ofcom has put off the auction for another two months. It seems the delay is coming from mobile operators or service providers who say they are worried about the connection with other auctions. They say they don't know how much to spend on 2.6GHz spectrum, till they know more about the availability of other spectrum. What's going on there?
Yes. We're interested in this, obviously. There's the 900MHz spectrum, and the UHF digital dividend.

Are they just trying to delay the auction to kill off WiMax operators in the UK?
Do people think that? I shouldn't comment.

Stepping back and taking an Ericsson perspective — how much of an issue is it for you, when that spectrum becomes available?
We have a very good roadmap and an ecosystem. The most important thing is that operators continue to invest in their networks — and have the spectrum that allows them to do that.

It's been suggested that since mobile data is usually done indoors, the LTE upgrade could be rolled out with small indoor 'femtocells' instead of traditional outdoor 'macrocell' base stations. Indoors, LTE handsets would get full data rates, over broadband backhaul, while outside, they would fall back to 3G. This could save operators money, and allow them to buy fewer LTE base stations from companies like Ericsson. Is it possible that the market for LTE macrocells might be vastly smaller than the market for 3G macrocells was?
I agree with some of this. Indoor coverage is an issue, and fixed backhaul has to be consistent with the capacity required for 100Mbps mobile broadband. We make an indoor base station, for 2G networks; 3G femtos and LTE femtocells may come along, but our focus is on 2G femtos because they work with existing handsets. Will femtocells delay macro networks? I don't know.