The wireless personal-area network technology is more than 100 times faster than Bluetooth, but business applications are still a long way off.
Ultrawideband wireless technology has been called "Bluetooth on steroids." Like Bluetooth, its personal-area network (PAN) cousin, UWB is designed to replace cables with short-range, wireless connections, but it offers the much higher bandwidth needed to support multimedia data streams at very low power levels. And because UWB can communicate both relative distance and position, it can be used for tracking equipment, containers or other objects. In a recent technology demonstration, Freescale Semiconductor, showed a UWB device that transmitted at a data rate of 110Mbit/sec at a range of up to 10 meters. That bandwidth is 100 times faster than Bluetooth and twice the capacity of the fastest Wi-Fi networks. It is enough to pump three concurrent video streams over a single UWB connection. Vendors are promising UWB products that support speeds up to 1Gbit/sec.
Waiting for UWB
While the prospect of 100Mbit/sec data transfers is exciting, UWB is probably three or more years away from widespread adoption, especially for business use, according to chip makers and analysts. Government regulators outside the US haven't approved the use of UWB, and standards bodies are arguing over the final specification.
Craig Mathias, an analyst at Farpoint Group, predicts that the first products with UWB chips, designed for home theater applications, will debut next year. Mass adoption of the technology won't come until 2007, he says.
Business applications, when they come, will center on UWB as a replacement for the Universal Serial Bus standard, says Ken Dulaney, an analyst at Gartner. UWB could be used to easily connect several laptops to a single projector to handle video or slide presentations, or it could be used to back up large files quickly, he says.
Eventually, workers could carry a portable storage device equipped with a system image and UWB connectivity. Users would be able to sit down at any workstation, connect via UWB and start working.
"It's very, very significant technology, and UWB is a guaranteed win," adds Mathias, noting that 50 companies are making UWB chips, including heavyweights like Intel. But vendors have yet to agree on a standard. Intel is backing one camp, while another industry giant, Motorola (through its Freescale subsidiary), is backing the other. UWB faces serious regulatory hurdles as well, "so it's hard for it to move forward," Mathias says. The US is the only country to approve spectrum for use by UWB radios. Regulators worry that UWB will interfere with a range of other wireless devices that operate in the same spectrum, including cell phones, says Steven Wood, a strategy planner at Intel.
Opposition by foreign chip makers and foreign governments might be lessening, however, says Gary Anderson, a delegate in the International Telecommunications Union (ITU). He says tests shown recently to regulators from the European Union demonstrated that interference is not a problem.
"UWB is here to stay, and within the next six months we'll see a great warming in the international community for UWB," says Anderson, who is CEO of Boston-based Uraxs Communications, a developer of UWB devices. The ITU is moving toward a UWB mandate that should be ready in less than a year, he adds.
Unfortunately, at the Institute of Electrical and Electronics Engineers, the matter of reaching a UWB standard is still "politically deadlocked," according to Wood and others.
The IEEE study group for the 802.15.3a PAN draft standard is at a stalemate as the two vendor groups push competing specifications. Intel is aligned with the 140-member Multiband OFDM Alliance (MBOA), which advocates the Orthogonal Frequency Division Multiplexing (OFDM) standard, while Freescale and the 30-member UWB Forum are pushing for Direct Sequence UWB (DS-UWB) technology. UWB won't progress until the two sides reach a compromise, something neither appears ready to do.
Martin Rofheart, director of UWB operations at Freescale, says DS-UWB has a "two-year time-to-market advantage" over MBOA approaches. Freescale has demonstrated a chip set called XtremeSpectrum, which it expects to appear in home wireless digital video applications this fall, he says. Freescale says the chip will support speeds of up to 1Gbit/sec over 2 meters and will be available by the end of next year. Intel's Wood says Rofheart's claim of a two-year lead time over MBOA is "ludicrous." Intel and other MBOA vendors will "make it a real contest" for DS-UWB vendors, and the MBOA has nearly five times as many vendors as the UWB Forum, he added.
Mathias says he wouldn't want to choose which approach is better or which will win out.
"It's still early, but the amount of innovation by all the companies involved is indicative of the great potential UWB holds," he says.
If UWB doesn't catch on, other technologies in the fast-moving wireless arena could take the lead. For example, a nascent Wi-Fi wireless LAN standard, IEEE 802.11n, is expected to offer bandwidth of 200Mbit/sec. Some analysts say it's possible that by the time UWB products arrive, 802.11n devices may be available, providing higher throughput than first-generation UWB devices. Although the technology most likely would require more power than UWB, one research firm predicts that it will leap ahead of UWB for home use. That might slow UWB's momentum for business use as well, but don't count on it.
While some vendors are already touting 802.11n compliance, that's "a clearly misleading claim," according to Gartner's Dulaney, who categorizes the standard as "embryonic."
Ultimately, the success of UWB will also depend on its cost. That's still an unknown, although chip vendors predict that volume prices will eventually drop to about US$5, the same as was predicted for Bluetooth. Yet the promise of Bluetooth as a universal cable replacement didn't come to pass because the benefits of adding the technology weren't seen as compelling enough to justify the incremental cost for low-end peripheral devices such as keyboards, mice and even printers. With its higher bandwidth, UWB may offer a more compelling reason to adopt wireless PANs in the enterprise, and eclipse Bluetooth in the process.
"Bluetooth is a loser," Mathias says. The Bluetooth Special Interest Group claims that the industry is shipping 2 million chips per week, but "who uses it?" he asks. "With UWB, the economic potential is so great that it's hard to imagine it won't move forward."