Ultrawideband renews high-speed wireless hopes

Ultrawideband (UWB) is shaping up as a high-speed wireless technology for the enterprise. But before its success can be assured,...

Ultrawideband (UWB) is shaping up as a high-speed wireless technology for the enterprise. But before its success can be assured, it must first navigate the complexities of industry working groups.

Specifications for UWB, including 802.15.3a, are being forged by a task group of the Wireless Personal Area Networking (WPan) working group, a subset of the IEEE.

The task group's charter carries a broad mandate that goes beyond streaming, encompassing what it calls "time-sensitive file transfers" such as media content.

Until now, UWB's main sphere of influence has been in consumer electronics, enabling applications to run multiple high-definition television video streams among several devices in a home or to display camcorder output instantly without cables, for example.

The technology's capabilities, however, have since caught the eye of the enterprise.

Using extremely short and broad low-power pulses to convey information, this flavour of UWB would operate at speeds of 110Mbps at 10m and 480Mbps at 1m.

At such rates, UWB could offer speeds equivalent to those of USB 2.0 or IEEE 1394/FireWire across short distances. Some analysts also speculate that UWB could develop into an alternative to Bluetooth in the future.

But even over distances of 100m, where speeds fall off to kilobits per second, UWB can increase Wi-Fi's efficiency by providing an access point with a user's precise location, allowing a smart antenna to "electronically direct the radio energy in the direction of the actual clients or users", said Chris Fisher, vice-president of marketing at XtremeSpectrum, a UWB pioneer and holder of patents on some of the fundamental technology.

In this context, UWB would, most likely, be a complement to higher-throughput 802.11 standards. Such standards promise 100Mbps shared bandwidth over a large coverage area.

To capitalise on this promise, however, the enterprise will have to wait until UWB emerges from its working group phase.

At the most recent working-group meeting earlier this month, the MultiBand OFDM (Orthogonal Frequency Division Multiplexing) Alliance proposal for 802.15.3a - backed by Texas Instruments and Intel - received 60% of the vote and is viewed as the leading contender for standardisation once a regulatory concern is addressed.

The group requires 75% approval to move a proposal on to the standardisation phase.

The OFDM proposal encodes data with the same standard used for 802.11a and 802.11g and divides the available spectrum into several bands that can be used simultaneously to provide interference robustness, according Stephen Wood, strategic marketing manager at Intel's R&D unit.

XtremeSpectrum and Motorola are backing the other major 802.15.3a contender. This proposal offers a more traditional approach to UWB, making use of the entire frequency band with a notch taken out for the 5GHz range to avoid potential conflicts with military frequencies.

The US Federal Communications Commission made way for the latest round of UWB proposals by formulating a rule last year that defined UWB and opened a large swath of unlicensed use with stringent requirements. More recently, the FCC expanded the definition of UWB, making the OFDM proposal possible.

The FCC's rulings could pave the way for the OFDM proposal to exceed the 75% approval plateau at the 802.15 working group's mid-September meeting.

Working group chairman Bob Heile said that a mid-September approval of OFDM could mean the availability of draft-based chips by mid-2004 and that the IEEE could fully ratify 802.15.3a by late 2004 or early 2005.

If the OFDM proposal does not gain the required 75%, it may open the door for XtremeSpectrum and Motorola, according to the IEEE.

In the meantime, XtremeSpectrum expects to demonstrate products based on its consumer-electronics chipsets at the International Consumer Electronics Show in January 2004 and plans to have equipment available - mostly for flinging data among plasma displays and home theatre equipment - before the end of next year.

The emergence of two slightly differing sets of standards has not escaped the attention of Julius Knapp, deputy chief at the FCC's Office of Engineering and Technology.

"We will move quickly to clarify any questions on how the rules are interpreted," Knapp said.

For example, both working-group parties agree there is ambiguity as to how a frequency-hopping UWB standard would be measured for compliance. Although an obscure point, one interpretation could tilt the vote against the Multiband OFDM Alliance by making it difficult to achieve the speeds needed by the 802.15.3a task group.

XtremeSpectrum has already acted to get clarification on this point, and the OFDM group will follow.

Knapp said the FCC is looking at the larger issue of measuring frequency-hopping devices.

"There's still an opportunity to review this in the future" regardless of the near-term outcome, Knapp said. "Something similar happened in the course of developing the 802.11a and 802.11g standard, particularly the 802.11g standard" with OFDM not qualifying under the interpretation of the rules at the time.

Meanwhile, it remains to be seen whether UWB will become a viable alternative to Bluetooth.

The 802.15.3a standard has a goal that exceeds anything that Bluetooth can deliver: a combination of low power requirements, high transfer rates and simplicity.

The IEEE has two other standards, 802.15.3 (the basis for the UWB variant, .3a) and 802.15.4, which may also help spell the end for Bluetooth. The former uses spread spectrum in 2.4 GHz to achieve 11Mbps to 55Mbps versus Bluetooth's 1Mbps; it has been finalised, but few suppliers are expected to adopt it.

The 802.15.4 specification, marketed as ZigBee, has the goal of offering very low speeds and long battery life: possibly just a few tens of kilobits per second but over periods of six months to two years, offering a replacement for infrared remote controls and alarm monitoring wiring. It may employ UWB just like .3a.

If UWB becomes the basis for 802.15.3a, it may become the most viable alternative to Bluetooth, although Intel's Wood said it would be at least four years before 802.15.3a-based technology achieves Bluetooth's current cost and availability.

But because of the proposals' working-group statuses, "you have a clear period in which Bluetooth has a reason for existence that just won't be contested," Wood said.

Glenn Fleishman writes for InfoWorld


Read more on IT strategy