Ericsson ditches Bluetooth
Fragmentation and US indifference
Analysis Ericsson, the inventor of Bluetooth, has ended years of aggressively positioning the short range standard for dominance and has stopped developing new products. This is a sign that the market has become one for volume suppliers, not innovators, but it highlights a broader sense that, while the current generation of Bluetooth will achieve a large base, the next generation will be pushed into niches by emerging alternatives —if they can avoid Bluetooth’s many mistakes.
Ericsson’s decision to pull the plug on its Bluetooth design and manufacturing activities do not sound a death knell for the short range wireless technology, but they do show that the standard has reached maturity – with no obviously viable next generation.
This means that innovators like Ericsson will turn to other technologies with greater market potential and Bluetooth, within a few years, will be confined to a few niches.
Ericsson spun off its Bluetooth group, Technology Licensing, which invented the technology, in 2000. It currently has 125 employees, 100 working in R&D and including the core of Ericsson's original Bluetooth development team. They are expected to be transferred to other parts of the company. Ericsson will still support Bluetooth chip customers such as Philips but will not develop new designs or seek new clients for the current one.
We should not overplay the negative impact of Ericsson backing away from Bluetooth. The standard is sufficiently mature that it can make its own fortunes without being too dependent on the support of even its most major backer. Just as Nokia’s ambivalence about the Wi-MAX standard it helped to create no longer has the negative effect that it would have done a year earlier, so Bluetooth has grown beyond its parent. But the move does show that the Swedish company has decided there is no future for the technology significant enough to justify the backing of a major player.
This comes after years when Ericsson has been highly active in promoting the enhancement of Bluetooth to be more attractive to its target markets. Having developed the initial concept of a wireless cable replacement a decade ago, the Bluetooth SIG was formed in 1998 and the technology was adopted as the basis of the IEEE 802.15.4 standard for low power, short range communications. However, despite that success, it was dogged by problems of interoperability, high prices and fragmentation, and never gained significant ground in the US - 65 per cent of Bluetooth shipments are in Europe, compared to 10 per cent in the Americas, and the less cellular-oriented culture of the US has focused most heavily on Wi-Fi for wireless communications.
The last example of Ericsson making major efforts to adapt Bluetooth and ensure its future growth came in July 2003 when it floated the idea of producing a lightweight version of the standard to compete with Zigbee in industrial control and automation. Jaap Haartsen, chief scientist at Ericsson Technology Licensing, had a team working on a version of Bluetooth that would use the same radio but optimize and shrink the media access controller (MAC) for real time automation applications, creating a derivative that consumed significantly less power. Haartsen said at the time that Ericsson had solved the main drawback of Bluetooth compared to Zigbee, its relatively high latency arising from its use of frequency hopping – frequency hopping that gives it better range and robustness than Zigbee’s ‘direct sequence spread spectrum’ approach.
There is no longer a role in Bluetooth chips for a company that expects to innovate and achieve high margins and the baton will pass to commodity, high volume chipmakers. This is a natural cycle that has also affected 802.11b, but the difference is that companies at the high end of the Wi-Fi food chain had the prospect of shifting their efforts, and R&D budgets, into new generations of the technology such as 802.11g and 802.11n. Bluetooth, by contrast, is looking increasingly like a dead-end as emerging alternatives such as UltraWideBand position to take its territory. As such, it will be a good market for volume chipmakers for a few more years but will not have an obvious future survival path.
The future for Bluetooth
One vendor leaving the camp does not kill a standard, even when that vendor was the key inventor of Bluetooth and, in the early years, played a role similar to Intel’s in WiMAX in terms of promoting its favored technology. But it is a good indicator of several trends for short range wireless communications. One is that, while volumes will grow - devices incorporating Bluetooth are predicted to quadruple in number between now and 2008, from under 100m to about 440m - success will lie in efficient manufacturing and strong OEM alliances, not in innovation. Ericsson is increasingly looking to intellectual property and advanced design for new revenue streams, and Bluetooth offers few opportunities there any longer. It said that there was no obvious business case now that so many companies were making Bluetooth silicon.
Another trend is for Bluetooth to be beaten back into certain markets rather than fulfilling its original, and wildly overhyped, promise to be a general purpose cable replacement in any applications requiring communications over short distances. Its key market is now, of course, the mobile phone, for connections to headsets, car kits and so on. Hence, having closed down its Bluetooth Technology Licensing group, Ericsson will continue to support products and offer Bluetooth software, but within its Mobile Platforms unit. It will also continue to be a promoter within the Bluetooth Special Interest Group, which drives the development of the standard.
The best market for Bluetooth is the mobile phone, with a niche, driven by companies like BT, for connecting handsets to the wireline network. Other key sectors are in-car communications and industrial control. All these are big markets, but Bluetooth’s position within them has now been placed into more realistic perspective. It is, after all, a low data rate, very short distance technology, and even before Wi-Fi went mainstream, it was naïve to portray it as a full scale wireless LAN. It can connect handsets efficiently to car systems, but it cannot support in-car internet access or transmission of multimedia to passengers as 802.11a can.
The next generation
That's fine, but the problem for Bluetooth is that newer short range technologies - its younger relatives in the IEEE 802.15 personal area networking family - will start to usurp even its true strongholds in the next generation. Current Bluetooth standards are gaining ground, but have not achieved a sufficient level of acceptance - especially in the critical auto market - to be irreplaceable; while the next generation versions may be too little, too late to fend off the challenge from UltraWideBand, the technology that could underpin both low and high data rate PANs within two years.
Bluetooth is not standing still and letting its rival take over by default. In June, the Bluetooth SIG took the important step of adopting Enhanced Data Rate (EDR) specifications which introduce new modulation approaches that boost data rates to 2- 3Mbps, rather than the current 1Mbps or less. EDR, importantly, also improves quality of service and the higher rates allow core applications to be achieved with lower power - more important in most Bluetooth markets, such as industrial control, than raw speed.
Last week, RF Micro Devices began sampling a Bluetooth system on a chip (SoC) supporting 3Mbps speeds, following market leader CSR, which sampled a Bluetooth core with SDR earlier in the summer, with full production scheduled for this month.
The Bluetooth Special Interest Group is already calling for input to a new roadmap to ensure a future for its platform. Earlier this year, the SIG set up a roadmap committee to work on directions and timing for the next major iteration of the standard, expected in late 2005 and called for any interested parties to submit ideas. This was widely seen as an acknowledgement that Bluetooth has often failed to listen sufficiently to the real requirements of its market, a factor in its slower than expected uptake. Markus Schetelig, chairman of the SIG, acting chair of the roadmap committee and a senior manager at Nokia, said: "Now that the key specs are mature, we are getting to the point where there will be a bigger payoff for being closely linked with market demands."
Also, two different SIG committees are looking at real-time streaming extensions for the standard - the A/V Working Group is developing a consumer electronics focused version, based on the Internet Engineering Taskforce’s Real Time Protocol, while a second group of PC-oriented companies is looking at streaming over IP.
Challenges from UWB and ZigBee
This effort is almost certain to be sidelined by UltraWideBand – whether or not it becomes an IEEE standard, the technology underpins several developments by chipmakers from Intel to Motorola targeted at digital media applications and will be incorporated in equipment from the likes of Sony from next year.
Even in Bluetooth’s core markets, it faces challenges from ZigBee and low powered versions of Wi-Fi. There is considerable pressure to unify wireless LANs and PANs, especially in the home, around common physical layers. Some argue Wi-Fi should be the common factor, and are pushing very low power implementations that would be suitable for Bluetooth-style applications and chip costs. Others favor UWB as the unified layer, and support proposals for the technology to underpin the next generations of both ZigBee and WiMedia, plus wireless implementations of USB and 1394 connections.
Of course, the newer standards will face many of the challenges that Bluetooth has. One is the interference question for any network in 2.4GHz space, a major issue for ZigBee. Now that ZigBee silicon is just starting to ship, research by InStat/MDR predicts that it will achieve shipments of 150m units by 2008. The biggest challenge for ZigBee, especially if it is to take on home applications such as intruder monitoring, as well as factory environments, is to ensure it does not suffer interference from other technologies in the 2.4GHz band, where Bluetooth and 802.11b/g also live. There is work underway on coexistence with 2.4GHz Wi-Fi, and some vendors are exploring use of the 915MHz band too, plus an UWB version would give greater spectrum flexibility.
Another danger to the newer standards, which should be flagged up by Bluetooth’s experience, is fragmentation. The political battles over the specifications for the UWB-based 802.15.3a standard – in which Motorola and Texas Instruments are fighting to have their approach ratified - raises the greatest question mark over the future success of UWB, and echoes some of the problems that have held back Bluetooth.
The Bluetooth community has often failed to unify its technologies to the degree required by the user base, and nowhere has this been clearer than in the car market, a natural for the technology, but which it has threatened to throw away with infighting and inconsistent products.
Car manufacturers have demanded that the Bluetooth community create a stable, standardized method for data synchronization between a car kit and mobile handset for applications such as telephone directory and express frustration that the Bluetooth SIG has failed to agree on a method. For instance, accessing a phonebook or SIM card via Bluetooth is demanded by customers but is still not supported in the current version of the Bluetooth Hands-Free Profile, according to BMW.
The wider issue of a standard set of profiles supported in all Bluetooth phones is also holding back adoption in cars. For instance, many phones support only one of the headset and handsfree profiles rather than both, which disappoints buyers who expect the ‘Bluetooth’ label to enable both types of usage. A separate phonebook access profile is under development.
As the Bluetooth SIG starts to address many of these issues – and as challengers are likely to be slowed down by their own efforts to do the same – the technology will get stronger before it fades away. But the best hope for Bluetooth is not to try to be all things to all people but to go for niches where it performs well, and for strong coexistence with its apparent rivals. It will be helped by the increasing ease of integrating multiple radios into relatively compact, low power devices such as handsets. Intel, Philips and Texas Instruments, among others, have dual-mode WLAN / Bluetooth systems and Finnish company BlueGiga Technologies has developed the first device that supports multiple Bluetooth modules and routes traffic between then and Wi-Fi, GSM, GPRS and Ethernet Lan.
Such developments no longer force device makers to choose just one technology and so Bluetooth does not have to win a head-tohead battle, just to provide a useful function alongside Wi-Fi or cellular. However, from 2007, we can expect this usefulness to be reduced in many areas, and for Bluetooth to be squeezed cruelly between Wi-Fi and the UWB-based group of technologies. Ericsson’s cooling of interest will not cause the death of Bluetooth, but it has highlighted its short life expectancy.
Copyright © 2004, Wireless Watch
Wireless Watch is published by Rethink Research, a London-based IT publishing and consulting firm. This weekly newsletter delivers in-depth analysis and market research of mobile and wireless for business. Subscription details are here.
Bluetooth is attack vector for mobile phones
Orange, Smart launch Bluetooth car
VIA backs Bluetooth
Second consortium unveils 'broadband Wi-Fi' proposal
Intel: WiMAX in notebooks by 2006
Intel invests in smart antennae to drive Wi-Fi, WiMAX
Nokia to rejoin WiMAX Forum
Sponsored: Benefits from the lessons learned in HPC