Ericsson floats idea for 'Lite' Bluetooth spec
Ericsson floats idea for 'Lite' Bluetooth spec
By Junko Yoshida, EE Times
July 7, 2003 (4:23 p.m. EST)
PARIS Ericsson Technology Licensing is considering developing a derivative of the Bluetooth standard that would compete against the nascent Zigbee specification in industrial control and automation applications-Zigbee's target market. The proposal has set backers of the two short-range wireless communications technologies at odds, and some critics said Bluetooth chip suppliers are only trying to insinuate themselves into a growing market with better margins than mobile phones.
Jaap Haartsen, chief scientist at Ericsson Technology Licensing, said the "Bluetooth derivative" would keep the current Bluetooth radio but optimize the media-access controller. He said that Zigbee-which will build on the 802.15.4 specification approved in May by the IEEE-may be optimized for automation but "at the expense of robustness." And while Bluetooth chips are coming down in price-shipments are expected to triple to 90 million units this year-Zigbee ICs have barely arrived on the market. The Zigbee Alliance expects to make a full specification available to its members in the second quarter of 2004. But Haartsen questioned the wisdom of pursuing another air interface such as Zigbee.
The idea of a new Bluetooth derivative, privately described by some in the industry as Bluetooth Lite, is still "at an investigative stage," Haartsen said. The Bluetooth Special Interest Group, which drives specification efforts, has not put the derivative on its agenda.
But some industry players have expressed interest in Ericsson Technology Licensing's proposal. Paul Marino, vice president and general manger of business line connectivity at Philips Semiconductors, called the proposal intriguing and said Haartsen had raised a "legitimate question" about the need for another air interface.
Joyce Putscher, director of converging markets and technologies at In-Stat/MDR, said a presentation Haartsen made last month at the Bluetooth World Congress, during which he raised the issue of an industrial-focused derivative, "did give significant food for thought on the merits of such a proposal. If a low-power derivative solution could be achieved that would rival the low power that Zigbee is reported to offer, it could present a viable alternative."
Others expressed some weariness with the Bluetooth Lite idea. "We looked extensively at the Bluetooth PHY [physical] layer while developing the 15.4 standard," said Bob Heile, chairman of the IEEE 802.15 Working Group on Wireless Personal Area Networks, who is also chairman of the Zigbee Alliance. The idea was rejected for a number of reasons, he said.
For starters, "it is extremely hard to build large networks with a hopper [a frequency-hopping scheme, as is used by Bluetooth], and large networks are an important part of the Zigbee application space," Heile said. "It is also hard to maintain network sync with a hopper without frequent wake-ups. That's tough on battery life." Further, "the hopper adds substanti ally to the time it takes for a new device to join the network." That may be "OK for Bluetooth, but it's not good for the Zigbee space," Heile said.
Even if Bluetooth Lite is ultimately embraced by the Bluetooth Special Interest Group, he said, the proposal offers "too little, way too late."
But the derivative may make business sense for Bluetooth chip vendors trying to extend their reach. Maria Khorsand, president of Ericsson Technology Licensing, said the Bluetooth community is eyeing the industrial control/home automation market as a hot growth area. Others described the industrial market as the most lucrative segment available to Bluetooth chip companies, which face tremendous price pressure in the consumer market.
"It is understandable that companies who have developed IP [intellectual property] for Bluetooth would like to use some of that for another application," said Nick Hunn, managing director for TDK Systems.
But he expressed some reservations. "It is always possibl e to take one standard and change it to try and fulfill another application, but the results are usually not elegant and often not cost-effective," Hunn said.
Pitches for the Bluetooth derivative are coming from silicon suppliers or IP owners, not from potential users, Heile said. "The suppliers want to see Bluetooth realize its original market number projections, and the way to do that is to hype additional applications whether or not they make sense. It sort of reminds me of the Bluetooth of two years ago," he said.
With or without a new derivative, Bluetooth has already inched its way into industrial applications. It is used in tests, trials and small deployments for horizontal industrial and medical apps. Putscher of In-Stat/MDR predicted that health care and manufacturing will dominate shipments in the next five years. Though these horizontal segments are not large today, their prospects for growth are "more significant long term," she said.
Bluetooth's biggest advantage over Zig bee today is that Bluetooth chips have achieved a level of maturity as commercial volume products. TDK has encountered "many companies that want to implement short-range wireless connections between their products," Hunn said. "In general these companies do not have RF experience or expertise and want to find a solution that works 'out of the box' as a building block."
While many wireless options exist, including low-speed RF modems, Bluetooth and eventually Zigbee, Hunn said that "for a company wanting to start today, Zigbee is not an option." It will take a few years for Zigbee products and interoperability to mature for mainstream usage, he said, adding that the problem is "not specific to Zigbee; [it's] a facet of any new radio technology."
In addition, Bluetooth has already been incorporated into a wide range of phones, PCs, PDAs and other systems that could be used to read or control devices in some industrial applications, Hunn said.
But Philips Semiconductors' Marino said ther e will always be a class of industrial applications for which Zigbee will be the more appropriate choice. "Not all applications are suitable for Bluetooth, because of the size of the network and latency associated with it," he said.
It's unclear how much Bluetooth and Zigbee would overlap in vying for industrial applications.
"I see no major differences in the system-level requirements" that the two can meet, Ericsson's Haartsen said, "although there may be some differences in emphasis for each standard." He said it is "doubtful" that industrial control and automation applications will mandate a completely new air interface.
The price points and robustness of the Bluetooth radio are the two main items Haartsen cited for proposing a derivative spec. The Bluetooth radio was designed to "filter and survive," he said. It results in "a slightly larger chip, but it's much more robust against interference." Compared with the direct-sequence-spread spectrum approach of Zigbee, Haartsen said, frequency hopping gives Bluetooth superior range performance. When faced with fading, Bluetooth may experience a throughput reduction, but Zigbee's signals could vanish completely, he said.
Heile of the Zigbee Alliance disagreed. The Zigbee protocol was designed to be "optimal for the control and sensor application space," he said. It is less complex than Bluetooth and offers superior power management strategies for better battery life. It supports a very large number of node networks for critical enterprise applications, such as industrial lighting, and incurs far lower latencies: It's capable of joining and leaving a network in milliseconds, vs. seconds for Bluetooth. It can switch from sleep to awake mode in microseconds, Heile said.
He noted that Zigbee was principally designed for the efficient transmission of short messages, Bluetooth for continuous voice communication without regard to coexistence. "Even though Bluetooth's raw data rate is four times Zigbee's, a data packet with a 40-b yte payload over the Zigbee network will take half the time it would take on the Bluetooth network," he said.
More important, Heile asserted, Bluetooth Lite would not be Bluetooth. The only common factor between the two is that both radios will use frequency hopping.
Indeed, Heile described Haartsen's call as "a proposal for a new protocol" that underscores an essential point on which he says he and Haartsen agree: "Bluetooth is not up to the task."
Copyright © 2003 CMP Media, LLC | Privacy Statement