Freescale's UWB soap opera
Confused? You will be!
By Peter Judge. Techworld | Published: 01:00, 19 April 2006
Martin Rofheart, director of the UWB division at Freescale, explained it all to us. But it all begins to sound like a soap-opera. In fact, his version reminds us of one 1970s sit-com. Soap is the story of two sisters, Jessica Tate and Mary Campbell, and the strange affairs of their feuding families. Ultrawideband Soap is the story of two rival technology groups, whose behaviour isn't much less unlikely.
The entire cast of Soap were dumb, incompetent, or just plain weird - except Benson, the sane and sardonic butler, who emerged at the end with a series of his own. In UWB Soap, Intel is rich, stupid Jessica Tate, Bluetooth is clumsy Bert Campbell, and - well, perhaps you can guess who Rofheart would cast in the role of Benson...
In previous episodes, Motorola spin-off Freescale, with DS-UWB, and Intel, with its MB-OFDM technology promoted by the Wimedia group, have been contending to provide a UWB basis for the next generation of USB and Bluetooth, with connection speeds up to 1Gbit/s over short distances.
Unfortunately, UWB has faced external opposition and internal squabbles. The mobile operators don't like UWB. No serious evidence of interference has been put forward, but it operates in spectrum the telcos have paid to license, and it gives phones yet another high-bandwidth I/O channel that the telcos can't bill for.
Meanwhile, Freescale's UWB Forum fought the Wimedia group in various standards groups, and lost three rounds out of four. The IEEE gave up on making a formal UWB standard, after years of blocking, but the USB Implementors' Group selected WiMedia for the next version of UWB, the European standards maker ECMA rubber-stamped Wimedia's specification, and the Bluetooth SIG selected it too.
So is Freescale washed up? On the contrary, Rofheart says its Cable-Free USB technology is about to mop the floor with its rival.
Confused? You will be after this week's episode of UWB soap.
Bluetooth's slapstick slip-up?
We asked Rofheart - are you disappointed that Bluetooth chose Wimedia, not DS-UWB? "On the contrary, it is quite possible that this choice by Wimedia significantly jeopardises Bluetooth as a high rate solution," he shot back. "You have to look across three dimensions - schedule, developmental timeline and relative performance."
The Bluetooth SIG's UWB adoption is a clumsy comedy pratfall worthy of Laurel and Hardy and a plank of wood, as Rofheart tells it. In one move, it's given itself big problems in all three areas, he says, by approving Wimedia, but insisting that the Bluetooth version of it should operate only in the 6 to 10 GHz frequency range. The FCC has approved UWB for use in the US, allowing it to use a much wider range, predominantly between 3.6 and 10.1 GHz.
"From the technology perspective, this is not just a small change," says Rofheart. "It needs new analogue RF silicon, and a new digital baseband. It's complicated by the fact that no one has taken measurements or done channel modelling."
He goes on: "It's going to be multiple years before there is a UWB Bluetooth product. The SIG doesn't anticipate product till the 2008 timeframe. Wireless USB will get there first by more than a year."
"UWB Bluetooth will be late to market, more costly and will be lower performance than other solutions," he says, predicting that the narrower band will cut the signal by 6 to 10dB, slashing performance by a factor of four to ten - a link that could deliver 100 Mbit/s at 10m, will only deliver 10 Mbit/s.
"This ensures that some flavour of Wireless USB will win," says Rofheart. Bluetooth and wireless USB do essentially the same thing - a wireless personal area network - and there won't be room for both. "A UWB version of Bluetooth will be late to market relative to the 'official' Wimedia-based Certified Wireless USB, and very late relative to Freescale's Cable Free USB." Fans of Soap might see echoes of the dumb guy, Bert
Why is UWB Bluetooth being hobbled this way? He hints at interference from the mobile operators: "The spectrum stakeholders have been opposed to UWB from the inception. They've tried to block it from the regulatory perspective, and then to force it up to higher and higher frequencies. Now they're trying to achieve that by other means." In other words, they're fighting crazy battles that don't exist. Soap had a character - the Major - who did that.
Intel's clueless upgrade?
But even if Bluetooth is walking into a closed door, how can Freescale expect its own wireless upgrade for USB to win out as a standard PC I/O channel, beating Intel's Wimedia, officially chosen for the Certified wireless USB upgrade?
That's simple. In Rofheart's view, Wimedia is also being stupid. As Freescale already told us, he believes Intel's Certified Wireless USB is the wrong upgrade, and is changing too much and will fall flat.
"Our view is that in personal area networking, the dominant solution today is USB 2.0," says Rofheart. "It delivers most value, and people use it. There's a billion devices in the market, and another billion in pipeline," and unlike IRDA and Bluetooth, most are actually in use.
"We are extending that," he explains. "That's what consumers want. Cable Free USB takes any USB 2.0 connection and allows you to extend it wirelessly. You can take a dongle and plug it in, or embed it in a mother board or miniPCI card. You don't need to worry about compatibility of the end connectors, because that's all solved."
Cable-Free USB is ideal for the kind of new devices that are being attached to USB ports now, he says. "A lot of the growth in USB 2.0 is in previously unconnected devices, for audio, video or other entertainment. It can be difficult to know where these connections are - wireless solves that."
When the first products arrive this summer - on the shelves in July, he expects - users will be able to have a wireless port extender that doesn't have to be connected to the laptop, he says. "It's a very simple first application. Consumers don't have to spend money on new PC, and all their legacy devices will continue to operate."
By contrast, he says: "Certified Wireless USB demands a redesign, with new hardware, software and a new protocol stack. If you want that to be a seamless experience, you will have to buy a new PC, otherwise you will need to get new cards, and download software." Souind clueless? Soap fans will remember Jessica Tate.
So Freescale is just being sensible?
"We believe the UWB marketplace will be decided by USB, and specifically USB 2.0, because that's the value of the 1 billion devices installed," says Rofheart.
The UWB Forum was set up to make DS-UWB a formal standard, and Freescale is ditching it because it failed, yes? Rofheart won't say that: "We are focussed on driving ourcable free USB in the market, and pulling out of every activity that is not laser focussed on cable free USB or USB 2.0." He does hint there may be a possible extension to a Cable Free version of IEEE 1394 Firewire later on.
So are formal standards no longer necessary? "Cable Free USB can publish its own standard," he says.
Will users and OEMs hesitate to adopt it? "We will see some market inhibition amongst both OEMs and end users," he admits. "But we will see broad adoption." Consumers, in particular, don't care if something is a standard, as long as it does something they want at a good price. "If I can buy an extender port with a dongle accessory, and it works, I see no reason not to do it - if it costs $29 or so."
So price will be very important? "Consumers don't buy technology, and they don't buy standards. They buy a user experience."
They'll go with what works? "Absolutely. Wireless models always go through their first few years as dongles."
But will OEMs want standards? "There are some OEMs that will not take the solution and some that will. The announced Cable Free USB products have been award winners. If you're delivering value to end users, customers will benefit, and that's where value arises." Thank you, Benson.
And the others are dumb?
Rofheart believes Freescale has a very comfortable lead over Wimedia: "It takes about a year from the time you can demonstrate a level of performance that consumer will buy, until the time you can hit a consumer shelf," he says. "It's not just getting the chips to work, but setting up a complete supply chain.
"As of this moment there has not been a public demo of Wimedia operating at that level. So their clock has not even started. It may be six more months, and then a year from there. Wimedia products are no less than a year away from the shelves, and may be as much as two."
By contrast, Freescale is ahead, he says: "We had public demonstrations last summer, and will have products in July. That's very much in the one year window."
He's not scared that Certified Wireless USB has the backing of Intel and Microsoft - and repeats his assertion that Cable-Free USB could extend USB 2.0 in much the same way as Wi-Fi extended Ethernet - even without backing from Intel or the IEEE. "We are staying true to Intel's original vision," he says.
"There's a precedent," he says. "Wi-Fi went against Intel's wishes. They originally backed the Home RF standard from Proxim, and only switched to Wi-Fi later, in 2001." This is true. The switch is reported here, and there are remnants of HomeRF on the web. But, strangely, the original www.homerf.org domain is now a wife swapping (no, not Wi-Fi swapping) site.
Don't miss the next episode
Like any soap series, there are cliffhangers. This summer, will we see Certified Wireless USB watching gormlessly as Freescale nips past it to the prize? Next year, will Bluetooth fall flat on its face? And will Freescale emerge the victor, raise one eyebrow and calmly serve the drinks? Don't miss the next episode of UWB Soap!