Dr Kevin Kahn is an Intel Senior Fellow and Director of the Communications Technology Lab. The lab is responsible for all communications technologies developed at Intel, including radio, optical, and copper physical layer technologies, CMOS communications circuits work, packet processing, and higher layer protocols.
Dr Kahn joined Intel in 1976 and throughout his 30-year career at Intel has worked in system software development, operating systems, processor architecture, and various strategic planning roles. He holds multiple patents in processor architecture and communications technologies.
Among Kahn's current areas of focus are broadband access to the home, wireless LANs and PANs, spectrum policy, and related Internet issues. Rupert Goodwins caught up with him at the Intel Developer Conference in San Francisco to find out what the future holds for Ultrawideband (UWB) communications.
Q: How do you see the state of UWB spectrum regulation in Europe?
A: It's better than I might have feared at one point. Japan's a bit different, but Europe and Asia are meandering to harmonisation, with the proposed rules in both places looking similar. You have a sensing technique between 3-6GHz [to avoid interference with existing services], and you're free above six. Our multiband stuff can cope with that easily.
The problem with UWB is that you make an enemy of everybody. Everyone who has a investment in an existing system can feel threatened and that they'll have to share their spectrum with a competitor who'll interfere. But a lot of the objections can be dealt with scientifically, some are on the fringes of what's measurable. Others are hard to see as a problem — if you had thousands of transmitters on balloons hovering above an antenna, perhaps.
So what about the objections from some network operators that UWB will increase the amount of radio noise, meaning they'll have to spend more money in building more base stations to ensure good signals?
Some of that has been quite naïve. It's really a fringe issue, stuff that happens right at the edge of the network, where you have to ask what's really going on that will be affected. You can always construct a worst-case scenario where there's a problem, even with current licensing, and it's really no different with new technologies. What's occurred because of the discussion process is a common agreement on what interference is and how to avoid it.
Also, people came to the realisation that nobody's interests are being served by being destructive, and that’s a lot better place to be than assuming that the bloke on the other side is an idiot or on a different agenda. The big players have an interest in everything working well. Intel, Texas Instruments and so on all have a big investment in current radio technologies and don't want to break any of that. It's far from being in our interests to screw any of that up. Our interests are aligned.
If this process of reaching agreement has worked with UWB, will it work with future radio systems such as cognitive radio [where radios will actively seek out efficient ways of using spectrum instead of being limited to preset modes and frequencies]?
TV white space is a perfect model of cognitive model debate — we'll go through the same argument again. If you talk to people who own spectrum in the TV bands, spectrum that you'd like to reuse, they are very defensive — no, they won't accept a transmitter on the same frequency. But then you say "How about 200 miles away?" and they go "Well, maybe". And then it's a matter of asking what it is exactly they want.
That's like "Will you sleep with me for a million pounds? Ten pounds?"
Exactly! Now we know what they are, it's a matter of haggling, of finding agreement. Radio is a very odd medium. There's a magic to it. Spectrum doesn't belong to anyone really — you can't impose man-made rules on it, it doesn't stop on physical boundaries. You have to generate acceptable probabilistic models that any particular solution is going to work.
The cognitive radio guys are going to end up in the same argument, a lot of people are interested in potentially much better use of spectrum. We have to do a much better job of using this resource. Once you do, there's a big net benefit. Everyone will do better. It'll mean breaking down some fiefdoms if it's going to work, but it'll be worth it.
What's going to happen after 802.11n?
There's a law of standards bodies — always look one step ahead. For 802.11, we are getting close to the limits of practical benefits and improvements, the costs/benefits of pursuing it will make it a tougher road to continue. We'll see evolution to make the best of new conditions so if the TV channels become available for opportunistic use, then 802.11 variants will come along that fit into that role. Or we'll aggregate different channels, so a network will say "There's 40MHz here and 20MHz at a different part of the spectrum, let's combine them." But at the moment everyone's consumed with 802.11n and 802.16e.
We can push down a linear evolution path for a little longer, then do radio systems solutions. So I'll say: "I have this existing method, I can change it and use it in a new way." OFDM is a case of this. We're not going to wind up with one single radio standard, but if they're close enough we can cover a flock of radio standards with one demodulator with knobs on. So we're working on getting as much reuse out of the platform as possible, evolving radio capabilities over time. But we're not narrow zealots for particular schemes. You'll always need different techniques for different solutions.
How about 60GHz? Do you see that happening, and if so when?
60GHz is an interesting place to go; I think it's the next stop for UWB because there's more bandwidth at 60GHz than UWB has and you can do some very interesting things with antennas to compensate for the way the frequencies are absorbed. As for when; these things have a natural lifecycle where everyone gets excited, then there's a period while the problems are worked out, a cycle of five, six or seven years. There are a few start-ups we're aware of and we're doing some work ourselves now, as our UWB work winds down I can see those people moving onto 60GHz. It's got its challenges, but it’s a good place to go.
