HP: How the mobile industry can do better

3GSM: HP's Phil McKinney says mobile operators are chucking away valuable data, and he believes software radios are just three years away
Written by Rupert Goodwins, Contributor

ZDNet UK grabbed some time with Phil McKinney, chief technology officer for HP's network and service provider business unit, at the 3GSM World Congress in Barcelona.

A man with many years of experience in the mobile industry, McKinney has no shortage of opinions about what's going well — and not so well — in the business. Find out why he thinks the mobile operators are throwing away 70 percent of their most valuable assets, and how they have to change to survive.

Q: Is there a lack of mobile innovation today?
A: The challenge isn't the lack of functionality. Devices have plenty of functions. The challenge is usability. Companies build devices like Swiss Army knives, hoping to get lots of tiny niches covered in one product. The next level of users, the ones we have to reach, don't want that. Mobile phones have the 80-20 rule, where you use 20 percent of the functions 80 percent of the time, but much worse. It's more like 95-5.

If you just try and extrapolate from what we have today, it doesn't work so well. Take our iPaq 6900. It's got five radios in it, it's the most radio-dense device in the world. Five, 10 years down the line, you can see these things having a dozen radios. So we want to have fewer functions, not more, perhaps just one radio per device connected to a hub that's got all the complexity in. That keeps the cost down and things simple.

What role will software-defined radio have in the future, and when will it arrive? 
Software-defined radio (SDR) is two cycles of Moore's Law away. There's a lot of work going on in converged silicon, with Qualcomm combining GSM and CMDA on one chip, for example, or Intel saying that 80 percent of Wi-Fi and WiMax circuits are the same, and that's going to be the thing for the next three years — this generation and one more. Then we'll have enough computing power to do proper SDR in the handset.

It takes a lot of power to do all that signal processing, but it's tempting. The radio side of phones takes the most research and development, but it's also the part that changes least. You design a new phone every year, but you don't want to have to redo the radio. SDR will be good for that: a tune-able antenna and some silicon and you'll be there.

We've already got SDR running on 35 base stations in the US, approved by the FCC (Federal Communications Commission), in conjunction with Vanu.

Have the regulators caught up with the technology yet?
There's been a lot of catching up, and mistakes. In the UK, the rules were changed for MVNOs [Mobile Virtual Network Operators], forcing a pricing system on the mobile network operators which meant they did no investment on their networks at all. They couldn't make the return needed. Ofcom has admitted that didn't work so well.

In America, there have been a lot of questions about what it is you're trying to regulate with SDRs. The FCC now approves the waveform, although the hardware is part of that — our SDR base stations run with an Intel Proliant server, so that's been approved as if it were part of the transmitter. If you want to use this system, you have to use that server, and Intel likes that idea! But there's been a lot of work on the Hill [Capitol Hill, where the US lobbyists roam] to make that happen.

What do you think of WiMax's prospects?
Sprint is going ahead in the US, and it says that it's worked out its cost per bit and will be profitable. But the key thing is customer adoption. How do you sell the product to the consumer? That hasn't been sorted out. It's like 3G on laptops. All our notebooks are 3G-ready, and we've even launched our first 3G product — that was announced with Vista. But it's hard to see how to sell them. In the shops, they're sold like other laptops, but they've also got a lot of mobile phone in them. Which way do you sell them?

How do you see mobile operators maintaining their revenues in the future?
They have to go and develop really hot mobile applications with partners. It's all about partners, and finding what users want. If you've got a million subscribers and 10 percent of them want to create content, create ideas, you have to let them. You're not going to get 100,000 developers any other way. You have to work out a sandbox for them to play in. And operators have to expose their IMS (IP Multimedia Subsystem) infrastructures to their partners, let them co-develop services. It goes against the way telcos think and have always thought.

How does Google work? It mines its users' data. That's how it makes its money. Mobile operators have more and better data on their users than even Google has. They know everything, when the user turns the phone on, where they go, how many calls failed because the other line was busy, how many failed because the network was busy. But 70 percent of this information they throw away. If it's not directly connected with billable services, they're not interested. Think what they could do with that. They've got to think of what they can do with that very valuable real estate they own, the mobile screen.

Are there many operators thinking like that yet?
One or two. It's not common. There's a way to go yet.

Editorial standards