Qualcomm’s heart may lie in cellular, but that doesn’t mean it won’t try to dominate unlicensed spectrum technologies too. Indeed, it needs to, since they will be so important in high growth areas like the internet of things (IoT), and the more protocols are in play, the more the US giant’s skills in tight integration come to the fore. Hence a string of acquisitions, of which UK Bluetooth pioneer CSR is the latest. It joins Airgo, Atheros, Wilocity and others in bringing Qualcomm the technology, expertise and IPR to influence standards in key markets, and so enhance the chances for future chipsets.
While the other companies mentioned on that list put Qualcomm in the heart of once-despised WiFi, CSR adds to its expertise in Bluetooth, a ubiquitous but relatively low value component of smartphones, but now set for a far more important role in the smart home and other IoT segments. There are various Bluetooth companies Qualcomm could have bought, but CSR has a long track record and a particularly important asset, its advanced development of Bluetooth Low Energy (BLE) mesh capabilities.
Some players – notably backers of the Thread Alliance’s would-be smart home standard – are promoting 802.15.4 solutions like Thread, 6LoWPAN and ZigBee on this basis, challenging Bluetooth’s bid to be the default short-range WPAN connection in the connected household. By acquiring the wherewithal to help BLE overcome this hurdle, Qualcomm is not only gaining some of the most advanced solutions in this sector, but possibly setting itself up as the godfather of a BLE-based smart home.
The company has already shown its readiness to lead, not just join, IoT-oriented standards efforts. Its AllJoyn technology, open sourced under the auspices of the AllSeen Alliance, aims to be a de facto standard at the peer-to-peer discovery and connectivity level of the IoT. This is competing with rival efforts such as the Open Interconnect Consortium (OIC), which is spearheaded by Intel (and until recently, Broadcom – see separate item). Now Qualcomm is reaching down to the PHY level, where it will be throwing its considerable weight behind BLE (as has Broadcom), rather than the 802.15.4 solutions. In both ZigBee and Thread, Samsung is the chip major which has shown the most enthusiasm to date.
In this context, the £1.6bn ($2.5bn) Qualcomm will pay, in cash, for CSR looks like a good deal, especially as the UK company has a strong reputation and product line which will contribute to Qualcomm’s bid to grab as many elements of the smartphone component count as possible (even reaching into RF now), and to extend Atheros’s reach in integrated home platforms with WiFi, powerline, Bluetooth and WiGig.
One of Qualcomm’s key strengths is its ability to integrate many functions on a single SoC and it both drives, and takes advantage, of the shift from ‘combo chips’ to full integration on the baseband, a shift which saves cost, power and space – all at a premium in handset chipsets.
Broadcom, for instance, has flourished in combo devices which include WiFi, Bluetooth, GPS and other options like NFC or FM radio. However, the company has failed to make the difficult but important leap to add cellular as well, and recently quit the mobile modem business it had briefly acquired from Renesas. CSR had also struggled with integrating its Bluetooth chips in either combos or basebands, but Qualcomm is unlikely to feel the same pain in this area and it will continue to squeeze as many capabilities as possible into the baseband.
On the WiFi front, it has squeezed combo suppliers like Broadcom and Marvell since 2012, when it started shipping cellular basebands with WiFi included. MediaTek and Spreadtrum followed suit in 2013 and since these are the world’s top three vendors of application processor/baseband SoCs, they have driven a rapid move towards WiFi being included on that platform.
Its new Bluetooth expertise will help Qualcomm here, though it is also developing separate combo chips of its own, because there is continuing demand for these in higher end devices which prioritize optimized performance over size and cost. From 2015, for instance, Qualcomm will offer separate mobile WiFi and combo chips to try to attack Broadcom among high end OEMs which still prefer an external connectivity solution (MediaTek already has an external combo). Linley Group estimates that combos will still account for 47% of the total WiFi chip market in 2018, even though most midrange and low end devices will have WiFi and Bluetooth on the baseband. Standalone WiFi chips for mobile devices are only 2% of the 802.11 chip market.
And of course, the real growth area will be IoT chips operating at ultra-low power, many running only BLE, but able to mesh with one another and link to IP-connected hubs such as smartphones.
These two key attractions – advanced mesh technology to feed into smart home standards, and new capabilities for the core mobile baseband/SoC business, show why Qualcomm wanted CSR. The US firm beat an offer from US embedded chip supplier Microchip, whose undisclosed bid was rebuffed in August as being too low. Microchip and CSR have been working on a BLE module together and the larger firm was understood still to be considering improving its proposal. Qualcomm will pay 900 pence per share, 37% above the close on Tuesday, and valuing its prey at 19 times projected earnings – according to Bloomberg calculations, this is above the average of 16.7 times, paid for chip companies over the past 12 months.
Though CSR had remained insistent that it had a bright future without a purchaser, this is clearly a good time to find a new owner with sufficiently deep pockets to ride the storms of competition in the wireless chip market.
Indeed, CSR chairman Ron Mackintosh hinted at that in his statement, saying: “While the CSR Directors believe that CSR is now strongly positioned to execute its strategy of delivering growth and sustainable returns in the medium and long term, we believe that the offer from Qualcomm provides CSR Shareholders with an immediate and certain value which is highly attractive. The CSR Directors believe the Acquisition recognizes CSR’s long term prospects and growth potential, and takes into account the dynamics of the global market and the competitive landscape in which it operates.”
CSR’s quest to find its natural place in the new wireless world has involved several restarts and several acquisitions or sell-offs. In late 2011, it bought location chip specialist SiRF in a bid to come out of its Bluetooth niche and vye with Broadcom and others in the world of integrated connectivity combo chips. But it acknowledged defeat in the mainstream smartphone segment when it sold many of its location and wireless products to Samsung in 2012. It also made an ill-judged move into image processing with the acquisition of Zoran, whose assets were also acquired by Qualcomm earlier this year.
Qualcomm said it would gain more “products, channels and customers” in two important areas, the IoT and automotive infotainment. These have been the most important target markets for CSR since it sold its mobile activities to Samsung and reverted to its Bluetooth comfort zone, after several missteps in trying to address a broader technology platform. Among CSR’s key products for the IoT space are the UniFi WiFi chip, which supports WiFi Direct, Soft AP and WiFi tethering as well as low power operation; and the CSR8811 and CSR8311 Bluetooth Low Energy chipsets.
CEO Joep van Beurden said in an interview: “You can see the strategic complement between Qualcomm and CSR – they are the global leaders in silicon for smartphones.”
Despite that obvious synergy, the mesh technology remains the crown jewel, especially as Qualcomm needs to establish the same influence over the smart home, and other IoT segments, as it has done in mobile standards and platforms. The BLE mesh is what turns CSR from a tactical product purchase into something more strategic for Qualcomm.
The firm has a long history of acquiring companies which have contributed inventions and IPR to key standards – from the relatively obscure, like Flarion, whose Flash-OFDM platform was the basis of the defunct 802.20 standards effort and became the basis of Qualcomm’s 4G patents arsenal; to the mainstream WiFi players. For instance, there was an intense battle over MIMO approaches for 802.11n, back in 2005, led by two WiFi specialists, Airgo and Atheros. Both, of course, are now owned by Qualcomm, which has also subsequently snapped up Wilocity, a major contributor to the 60GHz WiGig standard.
All these transactions show Qualcomm responding to the profound change in its market landscape, brought about by the slow but steady decline of technologies based on its CDMA inventions and IPR, and by the rise of new rivals from China. It has faced up to its
new realities, particularly that the modem, the main source of its unassailable power in 3G, is no longer the main driver of innovation and success, and it needs to have offerings in every aspect of the mobile and embedded chipset – connectivity, sensors, graphics, RF, and the software running above them (hence AllJoyn and many other developments within its open source software group). And it will not be enough to have products – a company with Qualcomm’s business model, driven by patents and being first to every new technology, must also set the standards, de facto or de iure, as it did in cellular.
This is why we believe CSR’s BLE mesh technology is the primary reason for the good price Qualcomm has paid. The UK firm has spent years looking for an application where it can be cutting edge again, and has found it in IoT-focused Bluetooth, with leading work on power reduction as well as on meshing. After experimenting with other markets, the smart home has opened the potential for CSR to leverage its unmatchable track record in Bluetooth again, and it has continued to innovate and gather IPR in the standard. Over a year ago, it announced a pre-standard implementation of the Bluetooth SIG’s adaptation of BLE for the connected car. When announcing its automotive-qualified BLE chips it said: “We won’t be able to wait for the Bluetooth Special Interest Group to come up with new solutions” but instead added proprietary modifications to the BLE spec.
Another area where it has led the field has been in enabling Bluetooth to mesh, something the standard was not designed to do. This capability became more urgent when the Thread Alliance was announced recently – spearheaded by Google’s Nest unit and Samsung, this is a credible attempt to seize the initiative in wireless PAN standards, especially in the smart home where BLE has such opportunities. The Thread specs are based on 6LoWPAN and 802.15.4, and would provide relatively easy convergence with ZigBee (if the ZigBee Alliance decides to play ball), squeezing Bluetooth significantly. The key reasons given by Thread for supporting a technology based around 6LoWPAN (though actually created by Nest) were its ability to support meshing and native IPv6.
There is now pressure, therefore, for the BLE community to accelerate its efforts in mesh, and CSR has the upper hand. The SIG admits it is only just starting to incorporate IPv6 and mesh into its standards, saying that “Bluetooth Smart Mesh is already possible, and we’re working to make it a reality under the Bluetooth standard”. In fact, CSR has that pre-standard implementation running, dubbed CSRmesh and installed in a smart lighting product with Bluetooth 4.0. It does not support full IP communications, but aims for lightweight data transmissions and low power, low cost devices.
Rick Walker, the marketing manager for IoT at CSR, told EETimes in a recent interview: “With CSRmesh, we have demonstrated that Bluetooth Smart can form the basis of mesh networks that can collect data from and control thousands of devices. The wide adoption of Bluetooth in tablets and smartphones enables rapid adoption and a simple personal user experience. CSRmesh uses the smartphone for direct control and doesn’t require a specific bridge hardware to be used to interact with the IoT devices.”
The pros and cons of full multipoint mesh (which ZigBee can support) can be debated in the IoT environment. Greg Ennis, VP of technology at the WiFi Alliance, blogged in response to Thread’s launch in July: “Don’t believe the hype about the purported reliability advantage of a mesh. Introducing a network detour adds a single point of failure … the ‘reliable’ mesh could in fact be twice as prone to loss of internet connectivity as a pure WiFi network.”
But the emphasis placed on the technology by Google’s proxy will make it more important for each IoT standard to tick this box, and supporters of IoT mesh argue that it helps address two of the key enablers of a broad and functional platform to connectms of different ‘things’ – increased range, and multiple nodes for many devices (other enablers, of course, include ever-decreasing power consumption, smartphone connectivity, low cost chipsets and real world applications). Lee Ratliff, low power wireless analyst at IHS Technology, commented: “We believe mesh technology is important for many IoT applications, especially those that require extended range or peer-to-peer communication.”
From Bluetooth’s point of view, adding mesh would not only remove a perceived weakness when compared to ZigBee and Thread, but would allow it to decouple itself from the smartphone or other controller, and be used in full systems rather than just accessories.
Mesh is just one example of how the standard is expanding itself rapidly to address the varied use cases of the IoT. With CSR holding valuable inventions and IPR in many of those areas, Qualcomm will inevitably enhance its position in the Bluetooth world, and the IoT as a whole.
AllSeen moves into smart lighting:
The Qualcomm-led AllSeen Alliance has created a smart lighting initiative, aiming to make its AllJoyn system the standard way for connected lights to communicate with other appliances, people and each other.
The idea is for any AllJoyn-enabled device to be able to control a light bulb – a smoke detector might instruct a light to flash when it is activated, for instance, or home lighting could be brightened and dimmed by security monitors.
The Alliance is developing the Lighting Service Framework (LSF), which will be available at the end of the year. It will provide application programming interfaces (APIs) to allow third parties to include lighting interactions in their applications, but its next task will be to persuade connected light bulb makers, and other smart home gadget suppliers, to support the technology. The most important company to attract will be light bulb giant Philips, which is involved in many smart lighting activities working with partners such as its former chip subsidiary, NXP.
One which is already behind the initiative is WiFi lighting firm LIFX, which has implemented AllJoyn and whose CTO Marc Alexander will chair the LSF committee. Other home automation specialists which have included AllJoyn in their products include smart home automation hub provider Revolv and baby monitor inventor Sproutling. Home appliance companies are signing up, including Electrolux, Haier, LG and Bosch.