Spec sprawl derails home net consensus

Spec sprawl derails home net consensus
SAN MATEO, Calif. — Wireless home networking is in such a state of disarray that any talk of a unified standard has fallen by the wayside. Indeed, no single home-networking scheme has what it takes to meet the consumer electronics industry's wish list: low cost, long range, little interference, high bandwidth, low power and interoperability with legacy products.

Reservations about muddying the waters with too many "standards" have evaporated. What's emerged instead is a wireless free-for-all, as two basic approaches — 802.11 and 802.15 — vie for dominance. Meanwhile, the short-range Bluetooth spec, which falls under the 802.15 personal-area-networking umbrella, appears poised for a market onslaught.

High on the agenda of chip makers, consumer electronics vendors and cable operators is ensuring that the warring specs meet some basic quality-of-service and security measures, two necessary features if wireless audio and video are to take off in the home.

Broadly speaking, wireless networking is divided into two camps, each with its own factions. The IEEE-802.11 wireless local-area network (LAN) camp appears to have the greater momentum, and proponents are seeking to extend its reach from coffee shops and kiosks into the home. But the wireless personal-area network group pushing the IEEE-802.15 specification claims it is better equipped to bring multimedia into the living room.

To show that its mission is bigger than Bluetooth, the 802.15 camp has just drafted a 2.4-GHz specification that offers 55-Mbit/second throughput, which it says is suitable for high-end video distribution within a home. Dubbed WiMedia, the proposal was crafted by a group made up of Broadcom, Intel, JVC, Kodak, Motorola, Sharp Labs and Sony. The spec is so new that most industry players outside the task group haven't even heard of it.

The group, which is promoting WiMedia's robust quality-of-service (QoS) and security measures, believes it can offer "802.11a's high data rates at Bluetooth prices," said James Allen, a member of the senior staff at Eastman Kodak Co. (Rochester, N.Y.).

The spec was created because "no other wireless standards can simultaneously distribute three different digital video streams, one Internet [connection] and three phones, and one CD audio stream perfectly," said Bob Heile, 802.15 working group chairman.

Proponents of the competing 802.11 standard, however, derided WiMedia as a glorified Bluetooth — a fancy way of getting rid of cable clutter around a TV. But Heile said the critics are mistaken. While 802.15.3 offers a nominal range of 10 meters, it's possible to change the power levels and take advantage of a high-link margin available in the standard.

"One can send multimedia data at 55 Mbits/s as far as 30 meters and at 11 Mbits/s for 77 meters," claimed John Barr, director of standards realization at Motorola Inc. and chairman of the 802.15.3 task group.

Heile stressed that WiMedia is not supposed to replace Bluetooth or any of the 802.11 wireless specifications for the local area. Just the same, none of the 802.11 specs are well-suited for high-quality audio/video applications, since they can't do things like beam three different pay-per-view movies to three TV sets in a home.

Proponents of WiMedia contend that consumer electronics companies are looking for a multimedia-specific wireless spec as speedy as 802.11 but less costly and power hungry, and simpler to implement.

The WiMedia spec threatens to undercut the momentum behind 802.11 wireless LANs, particularly the .11b variety. One of the hot topics at the recent Western Cable Show in Anaheim, Calif., was cable operators' growing interest in pushing the standard home.

The trouble with 802.11

"Right now, 802.11b represents almost the entire 802.11-based wireless LAN market. That is unlikely to change in the medium term and that is what players are betting on," said Navin Sabharwal, vice president of residential and networking technologies at Allied Business Intelligence Inc. (Oyster Bay, N.Y.).

The problem is that 802.11 was developed for data-centric enterprise applications and lacks inherent QoS features needed for home networking. Disciples say QoS will be worked into an offshoot specification called 802.11e, a media-access control (MAC) enhancement that will be applicable to the physical layer of all 802.11 offshoots. The spec will handle prioritization and low latency, both necessary for voice and A/V applications. It's expected to be ratified in mid-2002.

Even so, some warned that adding QoS extensions and patches to existing wireless LAN standards would enlarge the protocol overhead so much that it would become too costly to implement and would reduce audio and video throughput. When QoS features are applied to 802.11a, for example, its 54-Mbit/s data rate delivers multimedia throughput of 18 Mbits/s. In contrast, the QoS features of WiMedia can leave as much as 33 Mbits/s of throughput for multimedia.

"The 802.11 standards are great standards. But the problem is its own success," said Motorola's Barr. "While they try to accommodate more media-centric audio, video and imaging applications on the same wireless network, they'd still have to keep maintaining the network model and topology that belong to LAN and are controlled by LAN."

Unlike 802.11e, which guarantees a bandwidth based on bits rather than time, the QoS built into WiMedia offers guaranteed time slots for synchronous and asynchronous data streams with QoS provisions. And WiMedia has QoS support for ad hoc connections, something 802.11e lacks, Barr said.

Even as vendors and engineers weigh the merits of 802.11e and WiMedia, one group of chip companies believes neither is necessary. Philips Semiconductors, Cirrus Logic and ViXS Systems think they can optimize video at the system level for home gateways and client set-top boxes using existing standards such as 802.11b and .11a.

Philips Semiconductors recently tipped its plan to "transrate" 5-Mbit/s video streams down to a constant 2.5-Mbit/s stream for an 802.11b-based wireless home network. "We've designed a solution to make a video stream fit into an available bandwidth, so that we can build a wireless home network today by taking advantage of widely available, cost-effective, off-the-shelf 802.11 MAC and PHY [physical-layer] chips," said Mark Samuel, set-top marketing director at Philips Semiconductors. The company is also working on a proprietary QoS management scheme that makes it possible to set aside a bandwidth for video, he added.

Similarly, ViXS Systems Inc., a Toronto-based fabless chip company, is also said to be cooking up a wireless video-networking technology based on 802.11a that it will demonstrate next month. It operates on a 5-GHz frequency, using an orthogonal frequency-division multiplexing modulation scheme that is said to provide more spectrum, denser installations, faster speed and less interference. "We find it [802.11a] a sound foundation for extending home networking for video," said James Reinhart, senior vice president of product development at ViXS.

Physical-layer differences

The differences between the .11a and .11b implementations from ViXS and Philips may not be trivial. "There is a world of difference in what the physical layers of 802.11a and 802.11b can offer," Reinhart said.

Meanwhile, ShareWave, which was acquired by Cirrus Logic Inc., has floated a new protocol called Whitecap2, which can interoperate with legacy 802.11b-based products. A MAC chip running the protocol can also support multimedia QoS. "Whitecap2 provides much more robust QoS and link performance than 802.11e," said Kevin Daroca, vice president of sales for wireless products at Cirrus.

Still others believe that to do audio and video streaming right will require a move to a faster physical layer. After many months of contentious lobbying and voting, the IEEE passed last month a proposal for 802.11g, a spec that will boost data rates in the 2.4-GHz spectrum from the 11 Mbits/s now possible with 802.11b to a maximum of 54 Mbits/s.

The 802.11g standard "offers the data rate of .11a, while maintaining backward compatibility with .11b — without having two radios," said Matt Kurtz, wireless LAN product-marketing manager at Texas Instruments Inc. "To do A/V streaming over 802.11b is possible as long as you're the only one on the network. But .11b simply can't give you a high enough data rate without contention on the network."

Wired IEEE-1394 proponents seeking the path to wireless are taking a serious look at 802.11 developments. The1394 Trade Association's Wireless Working Group has jumped into 802.11e group activities, to "do a lot of explaining and education on what's needed for QoS tailored for asymmetric data traffic like audio and video," said Peter Johansson, a project leader of the 1394 working group. The 1394 camp is looking to add so-called "express traffic," which Johansson defined as "a mechanism for a prior allocation of time for access to a channel." This isn't a guaranteed access to the radio in the strict telco sense, but "we are asking that when a radio is available we get a shot at it," he said.

The lobbying effort appears to have worked. The 802.11e task group last month embraced the initial QoS concepts for A/V streaming proposed by the 1394 group. Acceptance enables the 1394 Wireless Working Group to develop a 1394 protocol adaptation layer for devices using the 802.11e QoS provisions.

None of the 802.11 standards will be able to take advantage of standards-compliant QoS until later next year when 802.11e gets ratified. Both 802.11g and 802.15.3 are shooting for the completion of their standards by the middle of next year.