ARM wins IBM analytics integration for its mbed IoT device platform

ARM has scored a significant goal for its mbed platform for internet of things devices and sensors, which will now be connected directly to IBM’s IoT analytics systems.

The two companies have announced plans to integrate the two platforms as the IT giant extends its tentacles into every corner of the IoT ecosystem via its cloud services, analytics engines and web of partnerships.

ARM is looking for ecosystem control of its own, though on a far lower layer of the complex IoT stack. Its mbed OS, launched last year, provides a lightweight operating system for embedded devices powered by low end processors or microcontrollers, and integrates with the whole mbed developer platform.

As well as unifying many types of chips around ARM’s designs at both processor and software/security levels – essential to keep up the defences against Intel – mbed was clearly designed to interface in a simple, standard way to the cloud-based big data platforms which will make sense of all that IoT-generated data.

IBM is one of the biggest catches in that respect, and devices powered by mbed-enabled ARM chips can now register automatically with the IBM IoT Foundation platform, which will analyse the information collected by those end points. Device makers also have the option to have the IBM software control their products remotely and alert users to events such as equipment failures.

As well as data analytics, the IBM IoT Foundation includes security systems to protect data from IoT devices, and access to IBM Bluemix, the firm’s PaaS (platform as a service), which handles and prioritizes data flow.

In a joint statement, the companies said: “This fusion will allow huge quantities of data from devices such as industrial appliances, weather sensors and wearable monitoring devices to be gathered, analyzed and acted upon.”

IBM’s general manager for the IoT, Pat Toole, said: “The IoT is now at an inflection point and it needs the big data expertise of IBM and little data expertise of ARM to ensure it reaches its global potential.”

IBM also unveiled IoT for Electronics, the first in a series of services based on its platforms, which it plans to launch for specific vertical markets. This one promises electronics manufacturers improved ability to gather data from the IoT sensors, which can then be combined with other data for real time analytics, providing instant insights into the clients’ production processes.

A meeting at IFA:

We spoke to IBM’s Bret Greenstein, VP IoT Solutions, and ARM’s GM IoT Krisztian Flautner at IFA in Berlin, to get a sense of the strategy behind the move. Greenstein said that this was a collaboration to create a solution that worked out of the box, with a clear go-to-market path for hardware developers looking to add data capabilities to their devices.

ARM’s Director of Corporate PR, Andrew Winstanley, pointed out that around 40% of all ARM shipments are now mbed-enabled, which means that the out-of-the-box market that the pair are targeting is really quite big. Greenstein added that the two meet in a strong collaborative middle ground – with IBM moving from the cloud down the layers, and ARM beginning at the bottom with the hardware and moving up.

Flautner noted that the move also helps to facilitate the shift in the developer community, from an area that previously only needed expertise in embedded design into one that must understand the web and how it interplays with the hardware. mbed aims to simplify this process, as well as ensuring that the end product is scalable and secure, two parameters that Flautner said are often ignored in favor of easy development.

Consequently, the mbed developer kits, and the IoT Starter Kit from ARM in particular, aim to provide hardware level compatibility with IBM’s analytics and software platforms. Greenstein said that the Bluemix application development environment itself, a mix of Open Stack and Cloud Foundry, provides the basis for the new abilities – with elements of IBM’s Watson AI being virtualized and run as software services, alongside IBM’s analytics platforms.

The end result is an environment in which a hobbyist developer can quickly set up a device and begin experimenting with the capabilities of the IBM cloud. As there is only a charge for these services once at scale and in production, the barrier to entry for the hobbyist is their expertise and the cost of the starter kit. Both ARM and IBM said that they are looking at simplifying the process for the developers – whether they are garage-based hobbyists or Fortune 500 engineers.

In a wider discussion about the IoT ecosystem, the pair agreed that the value is in data; with ARM facilitating the transfer of data from the network edge, and IBM turning those ones and zeroes into something valuable – or actionable, to use the dreaded term. Greenstein said that this data is only valuable for milliseconds, and that if it isn’t processed in time, it is wasted.

Our conversation then turned to how the IoT can increase operational efficiency, and the pair agreed with RIoT’s outlook that these technologies can be brought into existing business processes and improve the bottom line simply by reducing the amount of wasted data, time or physical assets, in a business.

Similarly, the data that is acquired can be tremendously valuable, especially to the engineering departments, as Greenstein noted. Once manufactured, an engineer has no idea how their product is used in the field. Consequently, BOM costs could be reduced by removing unused or underutilized features, and customer behavior can be monitored and used for design feedback. Once paired with predictive maintenance, to fix small problems before they become big ones, this platform creates a very valuable product for both the customer and the manufacturer.

But it’s not just individual products that stand to benefit from this increased connectivity and data capabilities. Flautner provided the example of a toilet cleaning company that can shift from a time-scheduled rota to a demand-driven schedule, where cleaners are dispatched in response to usage rather than an hourly rotation. Greenstein noted that time-based systems are arbitrary in an age where this sort of data insight is available.

Closing out the discussion, Flautner noted that the IoT is currently comparable to the internet in 1992, in that many are unsure about the particular direction the technology will move in. But, he noted, there is a key difference in that the IoT businesses know exactly what their technology is for, and what they need to do with it. Flautner added that it took around a decade for the value of the internet to translate into business models, via eCommerce and advertising, but the IoT is not a product of “hippies and DARPA” like the early days of the internet – these businesses have a vision, and the profit motivation to apply to their business models.