High-tech

Bluetooth: invisible networking, visible business?

By Ian Anderton

Bluetooth was announced several years ago as the technology that would revolutionise the way we communicate. In the lead up to every Christmas since 1999 we were told that Bluetooth gadgets - including intelligent washing machines and fridges - would be available for us to buy and that we would all be walking around wearing headsets.

In this feature, The Chilli attempts to separate hype from reality, looking into the challenges that Bluetooth has faced in its evolution and the progress it has made in product development. Also we offer a perspective on the question 'Is Bluetooth here to stay, and is there still scope for a decent return on investment?'

1. Background and history
2. False starts
3. Not all doom and gloom
4. Competing and complementary technologies
5. Opportunities
6. Conclusion

1. Background and history

The first usage models envisioned cable replacement technology and personal area networking primarily targeted at the PC and mobile phone markets. Then came the idea of using Bluetooth in an ad hoc network, with both peer-to-peer and multipoint connections (piconets). The number of potential uses for Bluetooth grew at a rapid pace: wireless printing, multimedia messaging, hands free telephony, Internet access points, telematics and so on.

Unfortunately there were also some outrageous predictions (billions of units shipped) made on its uses and delivery timescales, which gave Bluetooth some real credibility issues.

As early as 1994 Ericsson Mobile Communications undertook a feasibility study to develop a low-power/low-cost radio interface between mobile phones and accessories. 1998 saw the formation of the special interest group (SIG), the founding members being Ericsson, Nokia, Intel, IBM and Toshiba.

The first Bluetooth specification, version 1.0b, was published in early 1999. Even though this specification was somewhat incomplete it enabled early adopters, such as Parthus, Tality and TTP, to develop functional FPGA-based Bluetooth systems. These systems, and others developed using the Ericsson core, were still not ready for prime time. The 1.0b specification had too many inconsistencies to be used as the basis for viable products. Ericsson released its Bluetooth Development Kit enabling more companies to ride the Bluetooth bandwagon.

For the most part, 2000 was a continuation of the previous year's development efforts without any significant enhancements to functionality or demonstrations of real life applications.

2001, with the release of version 1.1 of the specification, saw the greatest advances, resulting in production releases of qualified intellectual property (IP) and limited product availability. Revenues from Bluetooth devices during 2001 were ~$70m. Some of the Bluetooth products sold during this time included 1.4m handsets, notably Ericsson R520, T39, T68 models and a Bluetooth battery from Nokia for the 6210. Cambridge Silicon Radio was one of the first companies to start selling silicon, integrating the entire Bluetooth product (excluding memory) onto a single chip.

2002 was marked by a shift away from unrealistic expectations to maturity of product with a significant increase in revenues and more widespread adoption.

2. False starts

What contributed to the delay in product availability?

i. The specification is both too complex, yet too loose. The core design document is over 1500 pages long, packed with detailed descriptions on each functional block with great emphasis given to timing, packet formats, link control and frequency hopping. Yet on the other hand there are several parts of the specification designated optional which are open to different interpretations or are not even implemented. Scatternet is a good example - it is overly complicated and does not have clear implementation guidelines. Few companies include scatternet in their list of features, and those that do are somewhat incompatible with each other. This tends to make it of little use, something that could have been pencilled in for later thereby reducing the development cost, risk and time to market.

ii. There was a significant delay between the start of Bluetooth development and the setting up of the qualification procedures. The rules governing the certification of Bluetooth devices - as governed by the Bluetooth Qualification Board (BQB) - were only established some time after the release of the specification and well into the development cycle. There were also not enough established test houses to meet the demand for product qualification. The lack of interoperability testing between different vendor products saw many products missing their market release window.

iii. Whilst the specification contains a significant amount of detail on each functional block that make up the Bluetooth device there is little attention paid to the interfaces that connect them together. This is especially true of the radio to baseband interface, creating a myriad of 'non plug n play' components. The protocol and profile software also suffers from this problem.

iv. The application usage model has been over complicated by the notion of profiles. Profiles are a set of software modules that control and set up the Bluetooth device for a particular application, such as a wireless printer, headset, serial port or LAN access point. A general purpose Bluetooth device would therefore need to have as many of these profiles included, while a specific Bluetooth device would require the relevant profile to be embedded. So what happens if two Bluetooth devices are trying to communicate but one doesn't have the right software needed to do whatever is required of it?

v. Interoperability with other wireless technologies and other Bluetooth devices. Bluetooth is a personal area network operating in an unregulated frequency spectrum that just happens to be the same as for Wi-Fi (802.11b/g) and ZigBee (IEEE 802.15.4). Consider what might happen at your workplace that has an 802.11 LAN enabled PC and integrated Bluetooth radio, Bluetooth mouse/keyboard all sitting next to your mobile phone. The lack of interoperability testing between different classes of Bluetooth products, as stated earlier, has created more delay and uncertainty in the buying decision. Gartner predicts that more than 80% of Bluetooth products offered in the next few years would have problems in basic peer-to-peer networking.

vi. Security. The adoption of Bluetooth into the mass market depends primarily on -four factors: an affordable cost point, satisfying a need, ease of use and confidence in its operation. Leaving aside reliability and interoperability, the inadequate security features in Bluetooth devices are causing some buyers to adopt a 'wait and see' attitude. This is especially so for businesses whose existence depends on secure data transmissions, and are not willing to carry the additional cost that may be required to take proactive measures against potential problem areas.

3. Not all doom and gloom

Despite the unintentional efforts of market analysts, and the technical challenges, the future for Bluetooth looks bright. There is now a clearer understanding and expectation of the role of Bluetooth in the wireless market space. The earlier stampede to the watering hole has given way to an enterprise-led effort to produce quality products and improve on the technology. Most of the speculators involved in the early days have been replaced by larger, more established companies, whose commitment and investment will guarantee the success of Bluetooth in the long term. What are the signs that point to this?

i. Bluetooth is making money, but won't for everyone. The general consensus (backed up by real shipment figures) is that 2001 showed a modest amount of revenue. 2002, despite the industry downturn, is on target to achieve ~400% growth in cumulative chipset shipments. Forecasts from In-Stat/MDR and Micrologic Research back up this growth rate by claiming that delivery of complete chipsets to product vendors (the complete Bluetooth chipset, not individual components) hit 35m by the end of 2002, with chipset shipments having a CAGR during 2001-2006 of 118%. The average cost of a Bluetooth chipset today is ~$10. Following announcements of single-chip solutions from both Infineon and TI, volume pricing is set to fall to around $4-5 during the 2003/2004 timeframe, as these manufacturers leverage their ability to integrate RF with logic on a CMOS process and manufacture in volume, knocking out many start-ups. The installed base of Bluetooth devices is projected to reach to reach up to 300m units by the end of 2004, fuelled primarily by the inclusion in mobile handsets and computing products.

ii. Bluetooth enabled products are becoming more readily available. Over the past 12 months there has been a notable increase in the availability of Bluetooth products, mainly in the three major mass-market sectors of mobile handsets, computing devices and consumer electronics. Sony-Ericsson, Nokia, Motorola and others began supplying Bluetooth phones earlier this year taking the major slice of a total 24m devices shipped. The number of Bluetooth mobile phones is expected to increase 3-4x over the next two years with the arrival of more players in the mid-low end market along with the pressure from the operators to increase revenues from the higher speed network services. Notebooks from Sony, Toshiba, Fujitsu-Siemens and Hitachi have included embedded Bluetooth controllers in some of their high-end systems. PDAs from Compaq (oops HP) and Palm can be transformed into Bluetooth devices by the addition of a PC card designed in house or from third party vendors such as 3COM, TDK and Socket. Xircom supplies a Bluetooth PC card adaptor for the Handspring PDA. Other products include digital camera/camcorders from Sony, Matsushita, Sanyo, wireless printers from HP, keyboard/mouse from Logitech and Microsoft, LAN access points from Red M, Siemens and so on. A full list of qualified products is available at www.bluetooth.org

iii. Windows XP and Mac OS X operating systems support Bluetooth. Apple have included Bluetooth support as standard in Mac OS X, Microsoft offer it as an update to Windows XP. Both companies offer limited support in terms of file transfer and synchronisation. Apple should be releasing iSync - software that allows data transfer to Palm PDAs and Bluetooth phones - shortly. Microsoft are promising Personal Area Network support at 'a later date'.

iv. Bluetooth is gaining acceptance in markets other than mobile phones and computers. The high visibility of Bluetooth in the mobile market has helped to create more opportunities in sectors that traditionally have not been a target for wireless products. The automotive industry has a need for Bluetooth to address particular issues with respect to the use of mobile devices within the car. Ongoing legislation in Europe and US prohibits the use of mobile phones within the car in 'non-hands free' mode - Bluetooth is a convenient solution for this.

4. Competing and complementary technologies

There are few alternatives to Bluetooth that address the needs of consumers to enhance their personal communications.

Cables are cheap, simple and work effectively. But they create unwanted clutter and do not lend themselves very well to freedom of movement.

IrDA has been around for some time and provides relatively high-speed connections. Unfortunately it suffers from the restriction on line of sight and is technically challenged on what it can do.

ZigBee is a low-cost, low data rate (max 250kbps) wireless personal area networking protocol targeted at automation and remote control applications, with the objective of having two AA batteries provide power for anywhere between six months to two years. Philips, Honeywell and Invensys kick-started the initiative, which is now supported by over twenty companies. It is also being standardised as IEEE 802.15.4. Its low data rate makes it suitable for control applications, whereas the strength of Bluetooth is its ability to support higher data rates.

802.11 provides the only credible alternative to Bluetooth, but its features and high cost do not lend itself well to the requirements of personal networking. 802.11 fits better into the longer range, higher speed access gateway and wireless hot spots. This provides an opportunity for 802.11 to dominate the more corporate environment whilst working alongside Bluetooth in the crossover to the more personal space. Click to read The Chilli's take on WLAN.

Similar technologies, such as WirelessUSB from Cypress, do not have the industry backing or versatility in application that could pose a threat to Bluetooth.

5. Opportunities

So are there still opportunities in Bluetooth? The window of opportunity for the enterprising start up to develop Bluetooth chipsets either as virtual components or fabless semiconductors, has come and gone. The list of suppliers for the hardware only is over 30 companies and includes the who's who in the electronic industry. The areas of development that offers most benefit are: combining Bluetooth with other complementary wireless technologies (802.11), addressing some of the technical challenges outlined above, creating Bluetooth products for niche markets such as gaming systems, medical devices or RFID type systems, where domain expertise in the end application is key.

On the software side, it's worth avoiding Bluetooth profiles, but opportunities might arise in developing protocol stacks that link Bluetooth into wireline systems, application software or even creating tools that simplify the build/integration process of Bluetooth systems.

6. Conclusion

As with most wireless technologies Bluetooth started off as revolutionary. It is now at a point where it needs to address some of the remaining challenges - namely those of interoperability, ease of use and security - whilst continuing to extend its utility to other growth applications and reducing system cost to stimulate the retail channel.

So, to the question: 'Does Bluetooth have any visible businesses?' The answer is yes, evidenced by strong revenues and the availability of products that consumers are purchasing. There are however, some key points that should be noted both by entrepreneurs and investors:

• Adoption time - investors must remember the time it takes for a new standard to be adopted. Think back to the adoption times for GSM, digital television and DVD. Consumer behaviour has a major part to play, especially with respect to safety, interference and quality of service concerns, necessitating consumer education to further develop the market.
• As described in The Chilli Value Test, start-ups should ideally focus on a niche that is being ignored by the larger players. In the case of Bluetooth, both TI and Infineon have targeted the market with sub $5 solutions, making full use of their ability to both integrate RF and baseband on a CMOS process and manufacture in volume. This could frustrate the attempts of early entrants to get a return on their investment through volume business.
• Standards setters should phase in the standards so that working (interoperable) products reach market quickly to give the new standard some traction and allow payback for the start-ups that bet on it
• Bluetooth is just I/O, and like RS-232, USB, will be integrated over time. By integrating the I/O (including RF) into a system-level architecture, new vendors can challenge incumbent vendors - Intel is using this strategy to establish its Xscale architecture, integrating I/O, memory, etc, against TI, Motorola and others in the mobile phone and PDA segments.

This article is derived from a research report available exclusively from The Chilli. Please contact us for further details at editor@theChilli.com

© Chilli Publishing Ltd 2003

23JAN2003