Ericsson banks on an HSPA future

world according to | Hakan Eriksson
LESS than three years ago WiMax emerged as a credible contender in the race to bring next-generation mobile broadband to market.

At the time Intel was backing the technology with promises to include WiMax connections in laptops, a mobile version of it was in the pipeline, and Australian taxpayers almost spent $1 billion using it to take broadband to the bush. The project was scrapped by the Labor Government, but last week Intel started giving manufacturers the option of building WiMax connections into their laptops for the first time.

The GSM Association has continued to pull the market in a different direction, promoting its 3G contender High-Speed Packet Access, which is based on the Wideband-CDMA standard.

HSPA has put 200 mobile networks around the world, including all of Australia's, on an evolutionary path towards 4G (or LTE - Long Term Evolution), which offers peak rates of 300Mbps. (The peak rate is the maximum amount of bandwidth it can share between mobile phones connected to a tower or cell at any time.)

People such as Ericsson chief technical officer Hakan Eriksson are the ones left to read the future.

He says the company has placed its bet.

What's your view on the current competition between WiMax and GSM mobile broadband standards? Is Ericsson agnostic about them?

WiMax will be around, but it won't be the mainstream mobile technology. If you look at the WiMax Forum's own prediction for mobile WiMax subscribers, its about 100 million worldwide by 2012.

At that time we believe that the number of mobile subscribers around the world will be around 1.2 billion, so WiMax will have a little bit less than 10 per cent.

As a consequence of that, Ericsson has decided not to develop mobile WiMax.

It would be wiser for us to spend our money on HSPA, but that doesn't mean WiMax won't exist.

Should Intel re-tool its manufacturing to add WiMax chips to its laptop systems?

That's up to Intel, and I wouldn't comment on their strategy.

I can say that we have made sure that we will have HSPA modules in basically all laptops that are shipped.

The GSM Association launched a new brand and logo last week for mobile broadband.

All the laptops that come with HSPA will have this sticker on them. You'll put the SIM card under the battery the way you will for your mobile phone and you'll have mobile broadband connectivity with roaming all over the world because HSPA is now available in more than 200 networks internationally.

There are about 800 devices from 100 vendors. HSPA is the established mobile broadband technology and WiMax will be on the side serving the niches.

I'm sure we will see CDMA, EVDO, HSPA and WiMax built into the laptops but HSPA will be 90 per cent of the world market.

When will these HSPA connections be available in laptops on a mass scale?

My laptop has it today, with a built-in module. It hasn't been available so much in Australia but in many other parts it is.

Lenovo met with analysts in Boston a few months back and they were very vocal about their ambitions to ensure that they will have this available in all their laptops.

Will the modules that you're supplying to individual manufacturers be able to support GSM technology that will come with LTE?

There will be modules that include both HSPA and LTE. Of course, they're not there today, but they will be part of modules that go into laptops next year. That is when LTE will start to be launched by the first operators.

How enthusiastic is Intel about embracing the GSM Association standard for mobile broadband as opposed to WiMax?

It's increasing. The reason for the interest in WiMax was a fear that there would be a big difference in patent costs for Wideband CDMA (upon which HSPA is based) compared with other technology, and that WiMax would be free of intellectual property costs.

Over time the industry has learnt there is not a big difference with the patent licensing costs.

That was one of the first reasons Intel was so interested in WiMax.

As that has become more and more understood there is less incentive to go for an alternative technology.

Has there been any indication as to when we might start to see LTE deployed in Australia?

No, it's a bit early to say when that will happen. There is a process to know what frequencies will be available and then carriers will have to claim them.

What about Telstra? Its Next G network is already on the LTE development path.

They could do that, but there's not really much point in doing so. The first movers going to LTE around the world - with the exception of DoCoMo in Japan, which is always early on new technology - will be CDMA operators moving off the CDMA track.

Nobody around the world is going into that. The big CDMA players around the world are going to LTE.

In other counties CDMA operators have left CDMA even earlier and gone to HSPA and that's what happened here in Australia.

There's no rush to LTE from HSPA - you're already on the right track so to speak.

How long will it be before Telstra will have to start moving down the track to LTE?

That depends on what they want to do with the network. The beauty of LTE is that it can operate in frequency-division duplex (FDD) and time-division duplex (TDD), so if you have spectrum that will only run TDD, such as 2.3GHz in many countries, the TDD mode of LTE is the best technology.

If you can run FDD, you can run HSPA and that is very good technology. If you want to go for really high speeds with peak rates up to about 100Mbps, that becomes more of a challenge with HSPA, even though you can do certain tricks with it.

Generally though, it becomes easier with LTE, even though it is no more spectrally efficient than HSPA.

If that's what you need, you have to move to LTE faster, but if you just want a high-capacity network with moderate to high peak rates around 10Mbps, HSPA will be fine for a very long time.

When will Telstra need to start increasing the capacity of its Next G network?

That's really a question for Telstra, but if you can acquire more spectrum you can increase capacity by running HSPA in more frequency bands.

If you want to launch an application that requires higher peak rates, you have to start looking at LTE.

There are different reasons for going to LTE, but HSPA is not becoming outdated.

Telstra is testing HSPA at 21Mbps downlink and it will probably have that before the end of the year.

From there it can go to 42Mbps and even to 80Mbps. With LTE it can go up to peak rates of 300Mbps. So there's still a lot of performance left in HSPA.

The Australian Government has set itself a goal of spending $4.7 billion to provide 98 per cent of the population with broadband at minimum speeds of 12Mbps. Most of the discussion assumes that will be done with some sort of fibre-to-the-node network upgrade, which will require extensive rewiring of the local loop and strand internet service provider assets. Is wireless technology developing quickly enough for the Government to deliver without having to fund the FTTN project?

Depending on what you have already invested in the network you will use different technology to reach the goal.

In many parts of the world there is already a lot of copper in the ground. With that investment already there, it makes a lot of sense to run VDSL2 so households can get speeds up to 30Mbps if they are within 500m of the local exchange.

If you want to go much faster, it starts making sense to dig down further with fibre and that's a good combination, which you see around the world.

In other parts of the world there may be no copper at all, and technology such as HSPA or LTE becomes attractive.

Could LTE fulfil the Government's goal?

Yes, you would be able to provide 12Mbps as an experience, but in a more densely populated metro area you would probably use a fixed system.

If you have a scenario where you want about 10 households to share an LTE cell, you could guarantee about 12Mbps steadily, but if you were in a residential area with lots of users, you would connect fibre and VDSL (very high-rate digital subscriber line).

For instance, out in the country you wouldn't take fibre to each farmhouse. You would put up an LTE base station and provide, say, 10 farms from that base station.

The Australian Bureau of Statistics says there is a sharp acceleration in take-up of wireless broadband in the past nine months. Do you have any ideas on what is driving that?

It's not just here. We're seeing that across the world. There are similar take-up rates across Europe and the Asia-Pacific region. On average about six times more data is being carried on 3G networks than on voice.

I attribute that to the convenience of having mobile broadband with you. People moved from fixed to mobile telephony for similar reasons.

On the fixed side, people went from dial-up to broadband and no one wants to go back to that because it's so slow and boring.

We've combined those two, so it shouldn't be surprising that it is such a success.

We're only seeing the beginning of this.

Is that making new fixed-line investments unnecessary and risky?

Again it depends on what you want as a service. If you want to start doing internet protocol television, you will need a fixed subscription in most cases.

In Europe, though, we see people giving up their fixed broadband because they realise they don't need to pay for both and they prefer to keep the mobile because it's more convenient.

At the same time, there are new services that will demand higher speed and they always come first on fixed lines, and later on mobile.