Bluetooth's Slow Dawn

Hype gives way to hard work as expectations for the mini radio network encounter nettlesome interoperability issues

10 min read
Ericsson T36, a triple-band GSM phone

Fig. 1(1) This Ericsson T36 is a triple-band GSM phone capable of working in the 900-, 1800-, and 1900-MHz bands. It is Bluetooth enabled, which means among other things that it can connect to the headset (shown here) without needing wires of any sort. In addition to Bluetooth, the T36 supports the wireless application protocol (WAP).

Last year, the people at Ericsson, Toshiba, Intel, Palm, and several hundred other companies were talking excitedly about the holiday selling season and the Bluetooth-enabled products they would have ready in time for it--cellular phones, personal digital assistants (PDAs), and notebook and laptop computers included. Today, only a few companies expect to have products ready for the coming holiday season, and most of them will be less technically ambitious--mere accessories for other products like headsets, remote control devices, and Bluetooth radios built into PC cards.

The key problem: interoperability, the root of all the enthusiasm for the technology. After all, few of these early devices have proven they can establish wireless contact and communicate with each other the way they are supposed to. To their chagrin, cell phone and PDA manufacturers who had hoped to begin delivering their Bluetooth products by last summer will be lucky if they can begin shipping by early next year.

Bluetooth was developed initially by Ericsson as a short-range (10 meters) cable replacement for linking portable consumer electronic products, but it can also be adapted for printers, fax machines, keyboards, toys, games, and virtually any other digital consumer application [Fig. 1]. Named after the 10th century Danish king credited with uniting the warring factions of Denmark and Norway, the technology provides a mechanism for forming small wireless networks of Bluetooth-equipped products on an ad hoc basis. It can also serve as a wireless bridge to existing data networks.

In what the Bluetooth community calls "unconscious" or "hidden" computing, Bluetooth-enabled products will automatically seek each other out and configure themselves into networks--most often, with just two nodes. Though small, such networks can be quite useful [Fig. 2]. They can forward e-mail received on a cellular phone in a person's pocket to the notebook or laptop computer in his or her briefcase; they can download data from a digital camera to a PC or cell phone; or they can alert their owners as they pass a Bluetooth-enabled vending machine. Bluetooth can serve as a means for connecting laptop computers or other devices to the public Internet in airport lounges and conference centers through permanent access points. It can also enable its user to exchange business cards with everyone passed on the street through a Bluetooth-enabled Palm--but not unless it has been given permission to identify the user to anyone or anything, which, according to a Merrill Lynch report on Bluetooth, "opens up whole new blind dating opportunities."

More than 2000 organizations have joined the Bluetooth Special Interest Group (SIG) and most of them are currently developing Bluetooth-enabled products under a specification developed by the group. The IEEE 802.15 Personal Area Network (PAN) Working Group, formed early last year, has made Bluetooth the foundation for a range of consumer network products, most of them portable. The PAN Working Group is currently developing a 1-Mb/s standard based on the work of the Bluetooth SIG, and hopes to define a "consumer-priced," 20-Mb/s or faster wireless personal area network that can be widely deployed for short-range information transfer.

The expectations for Bluetooth are huge. Analysts' market projections place Bluetooth-enabled products in the "next big thing" category, with sales expected to top a billion units by 2005. Motivated by those numbers, original-equipment manufacturers (OEMs) want to get their Bluetooth products to market as soon as possible to as wide a base of users as possible. But vexing technical issues have delayed the delivery of this much-hyped technology to the marketplace.

Interoperability blues

The principal problem is interoperability. Few of these products have met the Special Interest Group's requirements for interoperability or been tested under actual use conditions. One reason is that the qualification program was launched before conformance-test systems had been validated and were available. Not surprisingly, Bluetooth components are coming out later than expected and manufacturing capacity is limited. Another obstacle: Bluetooth products must be tested at a qualified test facility to ensure compliance with the Bluetooth specification [Table 1]. But no qualification test facilities are currently listed on the official Bluetooth Web site, as required by the Bluetooth consortium. True, some test equipment has been delivered with Bluetooth test features--Teradyne's MicroWAVE6000 instrumentation suite, for example. But no "official" or validated test hardware was available to manufacturers as of early October. Since interoperability is the be-all and end-all of Bluetooth, there is no point in coming out with a product until the spec is finalized or until some other means is in place for verifying that the item has a very good chance of working with other Bluetooth products when they become available.

Few of those involved are more frustrated than the Special Interest Group's Promoter Group. It was formed late last year by 3Com, Lucent Technologies, Microsoft, and Motorola, and joined by SIG's original founders--Ericsson, IBM, Intel, Nokia, and Toshiba. Its charter is to lead the Special Interest Group's development of Bluetooth technology and promote interoperability among members' products [see "What is Bluetooth for?,"].

Investments of resources have been huge--Microsoft Corp., Redmond, Wash., for example, reportedly has at least 60 people working on Bluetooth issues at least some of the time. But progress has been slow. "I have become an expert on worrying about Bluetooth interoperability," IEEE Spectrum was told by Mike Foley, a wireless architect in the company's Mobile Devices Division. To help speed Bluetooth products to market, the Interest Group has created what amounts to an interim qualification program with somewhat relaxed interoperability testing.

Tests of Bluetooth products and devices are now being done against designated protocol test products called Blue Units. These are based on development kits designed by the Cadence Symbionics Group, Cambridge, UK (part of Cadence Design Systems, San Jose, Calif.) and available through Symbionics and Sweden's Ericsson Microelectronics AB, AU System, and Sigma ComTec. The development kits comprise baseband and radio boards, interfaces (universal serial bus ports, phone jacks, and RS-232-C ports), host software, accessories, and documentation. They were created to enable first-time Bluetooth design engineers to accelerate the development of prototype products and devices.

But Blue Units do not carry the weight of full compliance with the Bluetooth spec; they are simply test beds to be used to establish confidence in key Bluetooth protocols. While they can test a number of key functions, their use is limited to partial testing of the baseband and link manager software. Blue Units cannot, for example, be used to test the logical link control and adaptation protocol, which adapts upper-layer protocols over the baseband; nor can it test the service discovery protocol, which Bluetooth units use to learn about the capabilities of other Bluetooth units.

Getting in step

The tests a product manufacturer has to pass to qualify a product for a Bluetooth trademark are divided into four areas:

  • Radio frequency qualification testing.
  • Protocol conformance testing.
  • Profile conformance testing.
  • Profile interoperability testing.

Protocols describe how Bluetooth-en-abled devices perform such basic tasks as service discovery, telephony signaling, and link management. Profiles specify which basic protocols and procedures are required for specific categories of Bluetooth devices and applications, like cordless telephones, headsets, and faxing. Profiles are the primary means for achieving interoperability among Bluetooth-enabled devices.

Currently available profiles cover cordless telephony, intercoms, serial ports, headsets, fax machines, local-area network access, file transfer, and dial-up networking. One group, chaired by Philips Semiconductor, Sunnyvale, Calif., expects to complete a written specification on three audio and three video applications by mid-2001. More than simply replacing cables for wireless speakers using Bluetooth technology, the group aspires to create applications that will make audio/video equipment interoperable with assorted consumer electronic products, as in moving data between a Bluetooth-enabled cell phone and a PC, or remotely changing channels and controlling the volume on television sets. Bluetooth accessories, such as PC Cards and USB dongles (security devices), are treated differently: they may be submitted for complete qualification testing only when they are installed in a host product.

Other working groups within the Special Interest Group are writing profiles for printing, imaging, location positioning, human interface devices, and personal area networks. Medical, automotive, and some other Bluetooth applications will require additional product certification.

Naturally, qualification requirements are expected to evolve as test equipment and procedures become available. In the meantime, designers must pass a series of tests based on four test reference categories established by the Bluetooth consortium. The categories range from merely informative, through self-testing, to full testing by a certified test facility using a validated reference test system. Since no validated test systems are yet available, the Special Interest Group is for the time being waiving the most rigorous category for manufacturers who pass the less strict of the categories and who successfully test their products against a Blue Unit.

Even that may not do the trick, though. "The Category A [most rigorous] test gives you a certain level of confidence that you will be interoperable with other Bluetooth devices," said Cedric Paillard, product marketing manager of Conexant Systems Inc. "It doesn't guarantee that you will be interoperable in the real world." Conexant acquired Ontario-based Philsar Semiconductor earlier this year to help speed its entry into the Bluetooth chip market. At some point, when the necessary tools are in place, Bluetooth products will all have to pass the most rigorous tests and also show that they are interoperable with what is called a "Golden Unit"--a SIG-designated unit that has passed all the required qualification and conformance tests.

Current plans call for continuing Blue Unit testing for a while after validated protocol conformance test systems have become available. Until then, the specifications remain the only authoritative source for compliance. At present it appears likely that RF testing to the level of Category A will begin at the end of December, with Blue Unit tests being removed some time next year.

Full compliance

As with all other RF devices, Bluetooth products will require regulatory type approval by international or national telecommunications regulatory bodies like the Federal Communications Commission in the United States. In real terms, Brad Tipler of Intel Corp., chairman of the Bluetooth Qualification Review Board, said, "The qualification process means it complies with the spec; it's a Bluetooth license. Type approval is a [government-issued] license to sell." Once a manufacturer has passed both tests, the item can be launched onto the world market with the Bluetooth trademark.

Because Bluetooth is still in its early development stages, test methodologies differ from those typical of more highly developed technologies. A Bluetooth application note published by Agilent Technologies Inc. suggests that "[Bluetooth] test procedures may require manual intervention or custom software control, as opposed to mature technologies in which easy-to-use, one-button measurements are available."

Part of the problem, according to Peter Cain, Agilent's Bluetooth guru, is that it has taken a long time for test houses to develop equipment with Bluetooth features, so that many Bluetooth designers have been forced to create their own test sequences and methods based on Bluetooth specifications.

Indeed, said David Lyon, chairman and chief executive officer of Silicon Wave Inc., San Diego, Calif., a Bluetooth chip vendor, "individual test houses have procured combinations of third-party gear and, in some cases, are using home-grown hardware and software to test both protocols and RF compliance." Several companies, he said, including his own, have built their own specialized testers in order to stress-test equipment under various combinations of protocol exercises while simultaneously monitoring RF parameters.

IC vendors and OEMs are using these home-grown test systems in so-called Unplugfests to test their products against other companies' Bluetooth products. Three of these events have been held so far, the most recent in August at the European Telecommunications Standards Institute in France. A fourth is scheduled for this month.

Unplugfest is no party

While they are not a part of the formal Bluetooth Qualification Program, the Unplugfests are very tightly controlled under nondisclosure agreements between participants. "We can't say where we are relative to others," said Matthew Phillips, marketing manager for Cambridge Silicon Radio Ltd., Cambridge, UK. "We can't talk about specific Unplugfest manufacturers and their specific problems."

In spite of the secrecy, it has become steadily clearer that finalizing the RF side of their design is one of the biggest problems facing would-be Bluetooth manufacturers. Indeed, Rangestar Wireless Inc., of Aptos, Calif., which is helping several companies develop Bluetooth designs, estimates that as many as three-quarters of the more than 2000 SIG members have little or no previous experience in RF or wireless design [see "Bluetooth basics,"].

Hoping to speed Bluetooth products to market, the SIG has formed the Bluetooth Measurement Initiative, whose task it is to work with test manufacturers to develop hardware and software for interoperability testing. In fact, third-party test equipment and software is now becoming generally available with preset Bluetooth tests.

Rohde & Schwarz is delivering what it says is the "first test system worldwide (the TS8960) specifically for Bluetooth equipment and components." Agilent Technologies introduced a Bluetooth DesignGuide in September, which is essentially an application layer that may be added to Agilent's Advanced Design System. It contains "system test benches" and reference designs, such as an optimal low-intermediate frequency (IF) receiver, for part of the RF portion of the Bluetooth physical layer. In October, Agilent also introduced a peak and average power sensor that covers the Bluetooth bandwidth.

Also, in October, Tektronix began shipping the CMU200 universal radio communications tester it announced in June. The CMU200 is a multi-standard test set for mobile phones with Bluetooth test features. Teradyne has shipped its MicroWAVE6000 instrumentation suite for testing Bluetooth radios and ICs to Ericsson, Oki, and Cambridge Silicon Radio, and has said it has several other customers lined up.

With these new test products now hitting the market, "the pace of testing should pick up over the next several months," according to Ron Wong, director of marketing of San Diego-based Widcomm, which produces the BlueStack protocol stack software. All the same, none of this test hardware has yet received "official" Bluetooth test status or been validated for Bluetooth testing.

Qualification grants companies the worldwide right to incorporate Bluetooth wireless technology in their products, and to use the Bluetooth trademark; but even the trademark does not guarantee that a product complies fully with Bluetooth specifications. Passing the qualification program demonstrates a certain measure of compliance and interoperability, but--as the Bluetooth SIG notes in its own literature--because products are not tested for every aspect of the Bluetooth specification, qualification does not guarantee compliance. In the final analysis, each manufacturer is responsible for ensuring that its products will interoperate with products from other Bluetooth manufacturers.

With several companies, mostly start-ups like San Diego's Silicon Wave and Britain's Cambridge Silicon Radio, all but betting the ranch on the success of Bluetooth, it is going to be difficult to soften the hype and face the reality of creating electronic products with an entirely new communications interface. But, then, this is supposed to be a joint effort and, as one market analyst put it, 2000 companies can't be wrong.

Key Bluetooth Web sites include:

https://www.bluetooth.com, the official Web site of the Bluetooth Special Interest Group (SIG). It offers background information and updates on Bluetooth technology developments and products, membership information, and copies of SIGnals, the SIG newsletter.

the site at https://www.topsitelists.com/bestsites/bluetooth/, which lists 20 Bluetooth company sites, with access to each site. The sites are ranked by "hits" per day.

https://www.mot.com/bluetooth/ Motorola's Bluetooth-specific Web site. It includes user applications, links to other Bluetooth sites, news, a calendar of events, a frequently asked questions section, and a glossary of terms.

Spectrum editor: Michael J. Riezenman

To Probe Further

A good book for learning all about Bluetooth is Bluetooth Revealed—An Insider's Guide To The Open Specification For Global Wireless Communications, which was published in October by Prentice Hall. The authors are Brent A. Miller, senior software engineer in the Pervasive Computing Group at IBM Research Triangle Park, and Chatschik Bisdikian, a research staff member with IBM's Thomas J. Watson Research Center.

Engineers interested in testing Bluetooth devices and systems are advised to get a copy of "Performing Bluetooth RF Measurements Today," a detailed application note from Agilent Technologies. Ask for Application Note No. 1333.

 

This article is for IEEE members only. Join IEEE to access our full archive.

Join the world’s largest professional organization devoted to engineering and applied sciences and get access to all of Spectrum’s articles, podcasts, and special reports. Learn more →

If you're already an IEEE member, please sign in to continue reading.

Membership includes:

  • Get unlimited access to IEEE Spectrum content
  • Follow your favorite topics to create a personalized feed of IEEE Spectrum content
  • Save Spectrum articles to read later
  • Network with other technology professionals
  • Establish a professional profile
  • Create a group to share and collaborate on projects
  • Discover IEEE events and activities
  • Join and participate in discussions