IBM's WPAN chipset aims to replace high-def cables, Bluetooth

San Francisco (CA) - The breakthrough bipolar CMOS (BiCMOS) chip announced today by IBM at the International Solid State Circuits Conference may be exactly what the IEEE had in mind when it convened task force 802.15.3c in March of last year. But the chip that would make it possible for high-definition CE devices to stream content between each other at gigabit speeds, may face an uphill challenge over the next three years, as IBM campaigns for it to become the backbone of the future wireless broadband home network.

At the new silicon germanium (SiGe) chip's official unveiling, IBM presented a remarkable chipset, compressed into a single package measuring 14 mm per side, yet including both the receiving and transmitting antennae necessary to exchange data at 60 GHz. Imagine, for a moment, the shortwave radio frequencies that bounce inside the ionosphere. The best antenna for a signal that travels the so-called "31-meter band," for example, is a straight copper wire extended 30 meters long. Its frequency averages about 9600 KHz. Now, imagine the size of the antenna that travels the so-called "millimeter bands" which IBM's new Wireless Personal Area Network (WPAN) chip deals with. In the company's demonstrator package, you can actually see the receiver antenna, suspended in space by a tiny metal rod, no longer than the letter "E" on the back of a dime is tall.

But signals at these extreme frequencies don't travel very far, and in fact, they're not meant to. This isn't really a WiFi chip, as even initial IBM reports led many to believe. The WPAN network is a dream of electronics engineers yearning to render obsolete the need for any cables whatsoever between home electronics equipment. But the next generation of high-def CE equipment is already slated to use, and may indeed be dependent upon, the next generation of high-capacity cables. There are already three such cable standards vying to reign in the industry, and in such a fracas, even the prospect of a wireless future could get squeezed out.

In an exclusive interview today with Brian Gaucher, research manager for IBM Research, who was with the IBM presentation team in San Francisco today, his company intends for this chipset to address the emerging needs of the consumer electronics industry. Manufacturers may, perhaps, be interested in deploying next-generation, high-definition streaming devices capable of transmitting signals over short distances - no more than 10 meters. Suppose for instance, Gaucher argued, that you have a high-definition display hanging on your wall, and a cable set-top box in your video cabinet. You need 1.5 Gbps to effectively stream content between those units. "How do you get rid of those cables?" he pondered. "That's what the consumer electronics folks are trying to target, putting out a wireless system that can throw those cables away, get rid of that, once you've [shelled out] $5,000 for the TV, the $100 cable set will look ridiculous. So they're trying to integrate all that stuff into a single solution."

The single solution the IEEE would like to see would drive 802.1 style architecture - of which current 802.11n wireless networking is one example - into the digital home. But since it's mostly engineers, not manufacturers, fueling this drive thus far, the IEEE ascertained in November 2004, the technology itself has to have already resolved the issues of how to address five key factors: attaining broad market appeal, achieving compatibility with other 802.1 architectures, attaining its own brand identity, meeting technical feasibility requirements, and at the same time remain practical to produce and market.

So IBM's premiere today is not just the introduction of another research technology, but the red-carpet rollout of a fundamental product concept: wireless multimedia components. For CE manufacturers to want to invest in it, perhaps the public has to become enthralled with the idea first. In the meantime, however, as Gaucher told us, WPAN connectivity becomes just another option for CE manufacturers to be considering: "I can't speak for what the HD community is going to do," he told TG Daily, "but [WPAN] just allows one more option." WPAN technology wouldn't change anything about how an HDTV or other high-def component operates; just how it communicates. So it's an easy choice to make, said Gaucher, if manufacturers want to make it. "But right now," he added, "it would just provide an additional option to all of the connectors that are there now."

Today, the global standard for personal-space wireless connectivity is Bluetooth, which some may consider WPAN's first and foremost competition in the computing space. But Gaucher believes the two are in different categories, by virtue of Bluetooth's less-than-broadband speed. "Bluetooth is a really slow technology," argued IBM's Gaucher. Its average throughput speed, he reported, is about 760 kbps, while IBM demonstrated 630 Mbps at the ISSCC conference today. The company's objective, he reiterated, is to meet IEEE's requirement of 1.5 Gbps, and then go beyond. "Bluetooth is stuck at less than a megabit per second, and if you wanted to think about Bluetooth in the personal area sense, one of the comments that gets made is that this 60-gig is kind of 'Bluetooth on steroids.' It's orders of magnitude better than what you see down on the Bluetooth side.

Could WPAN technology conceivably be integrated into a second-generation Bluetooth? Not from Gaucher's perspective; he sees WPAN's current competition as "the ultra-wideband folks who are pushing cable replacements for the 480 Mbps range, they're the ones who see this maybe, mostly as a competitor. But at least in the millimeter wavespace, I think we've got a lot of legs to grow, and we can take it farther in the future. Look what's happened with 802.11n, with this MIMO [Multiple In / Multiple Out] capability and spatial diversity...you start to apply those things to the 60 GHz band, it becomes even more interesting. Speeds could go up potentially higher than where we are today."

Following the IEEE's guidelines, and the timetable of the 802.15.3c committee, IBM intends to have its final WPAN standard complete and ready for production in the 2007-08 timeframe, Gaucher told us. This is right in line with what some are calling the "second generation of HDTV," whose components are already scheduled to use secure cable standards such as Intel's and Nvidia's existing DVI, the CE industry's HDMI, and/or (most likely "or") the Video Electronics Standards Association's DisplayPort. What makes these standards non-interchangeable with WPAN - at least for the time being - is that they are currently integrated with the security protocols necessary to prevent digital signals from being intercepted and duplicated without license. WPAN is just the transfer technology without the security protocol, unless the IEEE would like to consider adding a sixth requirement to its "Standards Development Criteria" anytime soon.