Networking and chip companies are racing to adapt technologies they created for the now-overcrowded mobile market, for the more open plains of the internet of things (IoT). Some have almost given up on the mobile device market in order to chase embedded gadgets in homes and cars, others seen IoT as a welcome extension to their business ‘ including, of course, Intel and ARM, which have taken their battle of the cores into the M2M world. As for the cloud and enterprise networking giants, the IoT presents a huge chance to scale their platforms from a relative handful of users to millions and even billions. The sector may be years away from that as yet, but several companies feel it is time to move beyond tentative product roadmaps and create entire divisions dedicated to capturing the IoT opportunity in its early stages ‘ including two of the biggest market makers, Intel and Cisco.
Intel has created a new Internet of Things group headed by Doug Davis as general manager, reporting to CEO Brian Krzanich and the executive office. That reporting line indicates the importance of this market to Intel’s plans to reduce its reliance on the declining PC space. Krzanich knows the extent to which the company can carve out new businesses will be the litmus test of his CEOship, and while he has made some decisive moves in the mobile space, he must know this will never be enough on its own, given its own slowing growth and the dominance of ARM.
Seizing a dominant role in IoT, where Intel has more differentiated missiles to hurl at its rival, should be more achievable ‘ and indeed, essential, if Intel is not to cede the device space entirely to ARM and retreat to the server space (which also, of course, sees opportunities in the data center and cloud platforms that must support mass-scale M2M data warehousing).
The choice of Davis also demonstrates continuity, since he is a veteran of Intel’s embedded business and was put in charge of the embedded and communications division in the chip company’s first really mobile-centric reorganization, in 2005. He ran the shortlived tablets and netbooks unit in 2011-12, but that was merged into the smartphones operation and he then did a stint running the wafer factory in Arizona.
Now he is back in his heartland, and some may see the appointment as a typical Intel move, staying with the people it knows rather than importing brave new ideas, but there is a real sense, at this early stage, in treating IoT as a huge expansion of existing frameworks and concepts, not a brand new technology ‘ certainly for a firm in Intel’s position, with considerable track record and years of R&D to put on the table.
The IoT Solutions Group will, significantly, include the Wind River unit. The real time operating system and virtualization software which Intel gained with that highly strategic acquisition is vital to its claim to be able to provide a credible IoT platform, and will be an increasingly important differentiator for the Atom and Quark SoCs. Previously, Wind River sat in the data center division, but it will be a great deal more impactful on Intel’s progress in the M2M space.
In the new division, there will be work to integrate Wind River even more tightly with the key low power SoC families ‘ the embedded iterations of Atom, and the new Quark range, announced this summer specifically for IoT applications. The group will also enhance these basic building blocks using technology from other areas of Intel, such as the McAfee on-chip security software and the Infineon Wireless modems and modules. The result will be a complete stack which could reduce implementation time, cost and complexity.
As Intel indicated when it unveiled Quark in September, it will spread its net broadly in terms of target markets for IoT solutions, though initial uptake is likely to be in those sectors which have been using and refining M2M for years, such as industrial control, and are starting to evolve from simple, low bandwidth systems to more sophisticated connected gadgets with greater intelligence and, in many cases, peer-to-peer links. Key markets will be home and industrial automation, retail, aerospace and automotive in the first instance.
As was clear when Quark was unveiled, Intel is excluding itself from some sectors and applications by not venturing into the ultra-low power, ultra-simple space, where ARM and specialized microcontrollers live. These dominate the existing M2M space, and ARM is powerful here with its M0+ Flycatcher core. Microcontroller-based SoCs are extremely small and power efficient and fall into the ‘silicon dust’ price range of less than 50 cents.
But while Atom will mainly power gateways and Quark will extend out to end nodes, these will be limited to those of a certain size and sophistication. Davis is keeping quiet about any plans to drive down into microcontrollers, which are not Intel’s expertise, but where it should at least secure a partnership in order to provide the end-to-end solutions some ‘industrial internet’ adopters will want.
For now, it seems that with Quark, and this new unit, Intel is trying to carve out a segment which it can define and control ‘ always a gamble, but necessary for a firm whose model revolves around leadership, not merely participation. The segment the company has marked out sits somewhere between a connected device that can run a full CPU and operating system, and one that runs a basic microcontroller.
Intel is banking on a middle market opening up between these two established categories, for devices which run an OS (embedded Linux) and CPU. It believes the rise of intelligence in M2M will drive the need for these more functional SoCs, which will therefore be able to bear slightly higher prices, power levels and size (and could also, in future multicore versions, be applied to low cost handsets). To address that hoped-for need, Quark is a whole new design ‘ not down at microcontroller level, but promising just 20% of the size of Atom and 10% of the power consumption. Building a dedicated division around it signals very clearly how high it sits in Krzanich’s priority list for building a new Intel.
Cisco has also been seeking to apply its traditional strengths to the IoT market, and has unveiled plans for routers and gateways to power what it dubs the internet of everything. Now, like Intel, it is putting these activities into a unit of their own, which can draw on technologies from around the group, while focusing single-mindedly on IoT solutions and on specific implementations for the vertical sectors which will drive this business. The new operation is headed by Guido Jouret as general manager.
Like Intel, Cisco will initially focus on sectors which are already heavy M2M users, but want to move towards broader, smarter and more complex systems in order to collect more data, improve efficiency and support new services. In particular, it will target manufacturing, oil and gas, transport, mining and other heavy industries, before turning its attention towards the more talked-about, but less developed IoT areas like smart home or smart cities. The new operation combines three smaller units, which had focused on smart grid, security, and vertical industries.
Cisco has been running IoT initiatives for about seven years now, and of course, this is a logical progression for the company whose switches and routers powered the original internet of people. Now the task is to adapt its switching, routing, networking, security and embedded technologies for this market and integrate them far more tightly to save cost, power and complexity. The new group’s seedbed consists of five product ranges for vertical sectors and these will be enhanced, even as their technologies are broadened to be applicable to a wider range of applications, in a more consistent manner. Key near term goals are to add more ruggedized routers, switches and gateways to the offering; to develop software that encapsulates legacy vertical protocols onto IP; and to seek to set standards.
While the actual solutions remain heavily Cisco-specific, the giant also wants to help bring standards to this sector. “The internet of things is a complex landscape and we saw there wasn’t a concerted effort to simplify it,’ Jouret said at a launch event in Barcelona last month. “There is a need to bring together technology companies, integrators and industry to accelerate adoption.” There are indeed many proprietary interconnects as well as battles over standards for sensors, connectivity protocols and other issues. In the key verticals this may not yet matter ‘ industries like oil and gas are well used to having their own technology islands. But as the IoT expands into mass consumer applications, standardized interconnection will become essential.
Of course, in Cisco’s ideal world, those standards would be based on its own software, which would then run optimally on its own hardware. It’s a trick the firm has pulled off before in some key aspects of enterprise networking and security, but certainly not all, and its place at the heart of the internet has always created conflict between the need for everybody to be able to interconnect seamlessly, and its ability to use its market strength to impose its own technologies and premium prices.
Where it scores either way is in its reach and range, enabling it to provide an end-to-end platform. Jouret told GigaOM: ‘Most people see this as a two-step problem with an endpoint and a gateway but you need a lot more than that,’ he said. ‘You need switching and routing and you may need a mesh network, or a ring or star topology for the data center. It’s not, ‘Let’s just put in some Wi-Fi’.’ Of course, another hallmark of Cisco is its readiness to fill any gaps via acquisition (one of the new unit’s functions will be to identify IoT start-ups and incubate new projects). Cisco has already purchased Meraki, which could help the firm move into cloud-based ‘IoT as a service’ systems, as integrators and equipment vendors start to pursue that model.