The Bluetooth short range wireless technology celebrated its tenth birthday at the recent Consumer Electronics Show. The standard has achieved a strong position in the mobile world, though in a very different role to that envisaged by its original backers, and while it once aimed to rule the personal area networking (PAN) roost, its future will now depend on how well it can integrate with other complementary standards. Bluetooth vendors face big decisions about how to define the next generations of their system to ensure it remains relevant despite the potential conflict with other, newer networks. And as the Bluetooth Special Interest Group (SIG) seeks to work out a role for the platform in its second decade, suppliers that are dependent on the technology also need to make key decisions about their future growth.
The Bluetooth SIG was founded in1998 around a short wave wireless technology developed by Ericsson and products hit the mainstream from 2000, since when 1.5bn devices have shipped. Although it was originally conceived as a PAN standard to connect PCs and other equipment to peripherals, in fact that role was largely usurped by faster wireline USB, which will almost certainly retain its dominance with the emergence of Wireless USB, and an expansion into consumer electronics and the digital home. Somewhat excluded from its initial target sector, Bluetooth nevertheless found a strong market in headsets and cellphone connections. Although it has consistently been portrayed as being in direct conflict with Wi-Fi, Wireless USB and other PC/digital home networks, in reality it has found its niche, and this has supported several specialist businesses, most notably chipmaker Cambridge Silicon Radio (CSR).
However, for companies like this to grow and Bluetooth to expand its reach, it is having to make some key decisions, most of which revolve around working alongside other technologies in the increasingly important and hotly contested next generation PAN. The most immediate priority is to ensure Bluetooth remains a near-default choice for multi-technology cellphone chipsets, with the focus on reducing power. An ultra-low power Bluetooth specification is imminent, according to the Bluetooth SIG’s executive director Michael Foley, and will enable functions such as wireless watches that can be used as remote controls. As well as offering a very low cost, low power option to device makers, this spec, which will be completed this year, could also take Bluetooth into some markets currently dominated by low power standards like ZigBee and ZWave, or by proprietary sensor networks. These include home or industrial monitoring and transmission of medical sensor data.
A major step forward in achieving ultra-low power Bluetooth was taken last June when the SIG decided to support the Nokia PAN technology Wibree (see Wireless Watch June 18 2007). As with the choice to support Wi-Fi and WiMedia, this was no doubt partly to ensure that Bluetooth works with its potential challengers rather than being attacked by them, but Wibree does add significantly to Bluetooth’s options in very low power applications.
Wibree was developed by Nokia six years ago and launched into the market a year ago. Although there were fears it would be presented as a vendor controlled alternative to Bluetooth, it works better as a complement, since it addresses devices with extremely low power capacity such as watches and heart rate monitors, and can allow them to compete with other devices such as Bluetooth handsets. Bluetooth is really too power hungry for such minute gadgets (it uses 10 times more power than Wibree), but a combination with higher powered cellphones could drive applications such as health monitoring for the elderly.
Wibree shares many of the characteristics of Bluetooth, with a range of 30 feet, 1Mbps data rates and operating in the 2.4GHz license exempt band. Like the older system, it is based on variable length packets that deliver low data rates. Early backers included Broadcom, CSR, Epson and Nordic Semiconductor. Nokia says that since Wibree shares much of the same underlying technology with Bluetooth it will be easy and cost-effective for manufacturers to design dual-mode chips. The SIG said in a statement: “Wibree consists of a small extension to a standard Bluetooth radio. As such, Wibree has the potential to enhance Bluetooth use cases in the phone by bringing sensor [network] type devices into the fold.” This hints at a long held ambition of the Bluetooth community – to extend its reach into the emerging market for automation and other networks based on tiny sensors.
All PAN technologies face the challenge of reducing power in order to live comfortably in the multi-network device that is becoming the norm, and to take on the biggest growth market for short wave wireless, sensor networks. They also face the parallel challenge of increasing performance to handle advanced multimedia applications, the driving force for all connectivity systems. The SIG is working on a spec for high quality video, a project that indicates how Bluetooth will seek to become ubiquitous by working with other technologies, not setting up in opposition to them. The high data rate version of Bluetooth will be built either on the WiMedia Alliance’s implementation of OFDM-based UltraWideBand, which is also the heart of Wireless USB; or on 802.11n fast Wi-Fi. It is likely that a Wi-Fi version will come first, as the 802.11n standard is close to stability, while WiMedia has been afflicted by various delays and performance question marks, but we see UWB ‘ whether WiMedia or another implementation like Pulse~Link’s CWave ‘ as being the main basis of all PANs in the next generation.
The Wi-Fi version is currently being presented as a stopgap solution until a UWB-based Bluetooth is practical. In 2006 the SIG and the WiMedia Alliance agreed to develop a high data rate solution that uses the unlicensed radio spectrum above 6GHz. The SIG envisages a convergence layer, like the one already devised by the WiMedia Alliance, that would be the basis of interfacing between the UWB MACs and Bluetooth’s logical link control and application protocol layers, supporting features such as service discovery as well as making Bluetooth suitable for a wider range of devices. Other PHY/MAC layers could also be plugged under the Bluetooth protocols in future too and with WiMedia and other bodies looking towards the 60GHz spectrum for very high speed applications, there may be SIG activity there too.
Before either of these new developments, the current version of the standard, Bluetooth 2.1 + EDR, will reach mainstream devices, with performance and usability enhancements. These include improved security features, advanced topologies and better quality of service, as well as multicast support so that data can be sent to multiple end points simultaneously, and eSCO air mode negotiation for renegotiation and seamless switching of the air mode.
Even here, the integration with other near range technologies is important ‘ one enhancement in 2.1+EDR is support for easier Bluetooth pairing. Among the pairing modes available will be the increasingly important NFC (Near Field Communications), the Sony/Philips standard mainly used for mobile payments and contactless smartcards (Sony recently integrated NFC with UltraWideBand for applications such as downloading music tracks to a phone while paying for them using NFC. Bluetooth could support similar combinations.)
Integration with other technologies may create a strong way forward for the standard, but it does not necessarily make life easy for the Bluetooth chip vendors. When Ericsson Technology Platforms exited the Bluetooth semiconductor market in 2004, it was a clear sign that the standard had reached maturity and commodity status, and that innovators like Ericsson would turn to other technologies with greater market potential and margin. The dominant Bluetooth chip suppliers are CSR, which has over 50% market share, followed by Broadcom with almost 30%, while other significant players are Infineon, Marvell, Philips, STMicro, Qualcomm and Texas Instruments. The major growth prospects for Bluetooth suppliers lie in increased integration of the radio into multi-technology systems, especially for handsets and other mobile devices. This has aroused new levels of interest in Bluetooth from the handset baseband experts, which had often previously perceived the PAN technology as too low margin to bother with. Qualcomm, in particular, is keen to see the cellphone taking on the role of home media server, and is packing functionality into its chipsets to suit. At the end of 2006 it acquired the Bluetooth assets of RF Micro Devices and is now reported to have bought audio streaming specialist Open Internet North America, which already provides a Bluetooth stack for the iPhone and Logitech FreePulse headphones. OINA also has its own lossless codec for high quality audio streaming, and has been experimenting with UWB.
Moves like this from Qualcomm put intense pressure on the key specialist CSR, which despite its huge success in Bluetooth has repeatedly failed in its efforts to diversify its revenue base into related technologies, despite attempts in Wi-Fi and other networks (see Wireless Watch November 10 2004). A year ago, CSR introduced a chip that combines the first version of its UniFi WLan technology and a choice of Bluetooth cores, claiming significant power consumption advantages over similar offerings from Broadcom, TI, Marvell and NXP. Dubbed the UniVox Mobile, the reference design, which uses the company’s tried and tested software stacks, ensures coexistence between both 2.4GHz radios with Collision Detection Logic. However, it has yet to make much impact on the market, and in the meantime the baseband providers are leveraging their established relationships with device makers to ensure they steal a march on CSR in multimode products. This is particularly noticeable at Qualcomm and Broadcom. The latter is poised to steal the market lead from CSR on the back of its appointment by Nokia as one of four key handset chip suppliers, with many of those phones likely to use an integrated EDGE/GSM/Bluetooth/GPS solution, edging CSR out of many Nokia deals. Broadcom and TI have incorporated more options into their multi-radio offerings than CSR (including FM for instance) and symbolic of CSR’s dilemma is the fact that its Bluetooth chip is included in the current iPhone, but the upgraded Apple device is highly likely to incorporate the Broadcom Wi-Fi/Bluetooth module instead from this summer.
For the standard and its suppliers, then, integration is the name of the game, and while the SIG seems to be working effectively alongside the Wi-Fi Alliance, WiMedia Alliance and other key players, market leader CSR has a tougher road ahead.
Bluetooth facts:
Bluetooth was invented at Ericsson and named after Danish king Harald Blatand, who was nicknamed Bluetooth because of his love for eating blueberries. He also united warring factions across Scandinavia, and so seemed an appropriate name for a technology aiming to unite previously unconnected devices. The Bluetooth logo is a runic symbol of Harald’s initials.
The membership of the SIG has risen from five to 10,000 companies, with the newest addition being headset maker JayBird Gear.
The most common application of Bluetooth technology is to connect mobile phones to headsets or cars, followed by transfer of data from cellphones to other phones, PCs or printers.
Newer applications include the incorporation of Bluetooth in jackets, motorcycle helmets and sunglasses with built-in wireless headsets, and gaming controllers.
Ericsson spun off its Bluetooth group, Technology Licensing, which invented the technology, in 2000 and exited the market in 2004 (see Wireless Watch September 2 2004), at which time the group had 125 employees, 100 working in R&D.