At stake is a market potentially bigger than Wi-Fi, with UWB touted by many as essential to consumer electronics -- doing for that what Wi-Fi's done for home computing. But many questions remain: whose technology best matches the way people will want to use it, which one will best match whatever the regulators allow, and how bad the risk of interference to other services is.
John Barr, director of standards realisation, and Martin Rofheart, director of UWB Operations, both at Motorola, spoke to ZDNet UK about recent developments in UWB.
ZD: How do you react to Intel's statement that it will push ahead with the MBOA and go to market regardless of the standards committee?
MR: Direct Sequence will be in the market and it will be there first. We have second-generation production silicon, and third generation coming out soon. Intel has PowerPoint presentations. We've shown three streams of high-definition video simultaneously at our booth in CES with Samsung.
JB: Intel would like to become the standard, but they can't convince the wider technical community because it's full of people who know what they're talking about. You don't create a standard just by putting up a Web site with 70 logos on it. But they own Wireless USB and are trying to pull UWB back into that. They're trying to block any other standard, that's the way they've been operating in 802.15.3a from the beginning.
Is it going to have to be one standard or the other?
MR: We share common spectrum with the Intel solution, so the sensible thing is to have harmonisation and assure quality of service. At the last meeting [of the committee] in Vancouver we did what we could to come to an agreement. We made a proposal so that we can come together -- come up with a standard that works better than just picking up one of the standards. And then at the last ad-hoc meeting, we had good participation to move the standard forward by letting different UWB radios talk to each other. It isn't costly to do this -- you don't need extra transistors and from a technical perspective there's virtually no cost.
JB: This common signalling mode will let piconets [a small network of devices] talk to each other regardless of their signalling mode. This helps with worldwide issues of regulations, performance and scalability. If you want to have a short-range, fixed architecture Wireless USB link you can do that, but other devices in range can choose higher speed, shorter distance or slower speed, longer distance links at the same time. It makes for much better power use and range of options.
The details have yet to be decided, but the common signalling would be a very robust 10 megabits a second, which would add less than 1 percent overhead to a link. Several of the CE [consumer electronics] companies in the MBOA have said that this makes a lot of sense, especially as there's no proof that everything people are saying today will work as promised.
With the standards process, we'll produce a report, get other companies to report on how they'll build on that compromise, and create a proposal that includes the common signalling mode. We don't know what Intel will do -- they're asking us questions about direct sequence, about some of the issues they've concentrated on. We'll spend time talking to Intel if they'll listen to us, and time talking to the CE companies in MBOA. If Intel is really interested in creating a broad market for UWB they can't walk away.
The only reason would be to lock the market down and force people to live with their solution, poisoning the market for the wider range of solutions. Companies like Apple are saying "we want to use it for something like an iPod" -- they want a total solution, not just a piece. The companies in 802.15.3a have not been offered an alternative for a permanent solution. Now they have one. Keeping this in the IEEE keeps it open -- places like Japan are worried that the MBOA are taking the standard away from the IEEE, and Japan wants to be involved with this to create a global standard.
MR: In my view, Intel has been posturing -- when we start shipping, they'll have to listen. It's been cast as a standards battle between Intel and whoever, but it's a race to get a solution to market first. What any organisation wants to do is achieve through standards or a special interest group what it can't achieve in the market alone. When Intel might say there's no industry support [for the Motorola proposal], how can they explain that five meetings in a row didn't move the needle from 60:40? People who don't want to fight with Intel in public will fight to keep a path alive in private.
JB: CE companies like Samsung, Sharp and so on don't want a Wireless USB solution. They want a solution that fits their space -- peer to peer, portable devices talking to fixed devices, mixed modes, and a very good quality of service -- which Wireless USB hasn't demonstrated. Certain products will take Wireless USB, but not things like digital cameras and hard disk-based portable solutions like the Apple iPod. Wireless USB is too complex and power-hungry. Intel and its companies look at UWB as a PC standard.
MR: Motorola is a believer in direct sequence, because UWB is a technology that's totally different from any other wireless. Most wireless is low bandwidth and high power -- making it wideband, which is what MBOA is doing, is expensive and complex. With direct sequence, bandwidth can go up and up with simpler, lower cost radios.
Both MBOA and direct sequence can do video to a high-definition screen in a room. We say we're better, MBOA argues that it is. But we can do low-power, hand-held applications -- 500 megabits a second to a gigabit a second and above, over one to two metres. MBOA has a fixed bandplan, and lots of complexity comes in then. They can't increase bandwidth except by adding complexity and power consumption, and we're the exact opposite.
How about regulatory issues?
MR: There are a whole load of questions about MBOA, and none about direct sequence. The NTIA [National Telecommunications and Information Administration, the US advisory body on spectrum use] is doing a whole load of tests at the moment, and they'll have results in three to nine months. Both sides have done tests, both have done their analysis, but both are saying opposite things.
JB: The FCC rules [that Intel claims compliance of] say things like 'stop frequency hopping to check your noise' and Intel has said that this isn't applicable to them. The FCC is saying that the rules are one thing, but there must be proof of non-interference. Intel is saying that because it complies to the rules, it doesn't interfere. That's not right.
Intel has published details of tests with digital satellite receivers that it says prove it doesn't interfere...
JB: We've got the identical test rig and get the opposite results. We ran the test with everything set up as it would normally be; Intel set things up on the edge of failure. Who's right, who's wrong? The NTIA is using their own test setup and will have nobody there to fudge the numbers. This sort of thing is very common from a regulatory perspective. It happens all the time.
What happens if Intel presses ahead with its technology and the NTIA proves it doesn't interfere?
MR: If Intel is right and has product out and nobody cares, then this discussion is moot. But if the Intel/TI analysis is wrong, then it's hugely impactful.
They're betting that the Motorola solution will go away. This won't happen. This never happens.