This past week I got the chance to visit the ARM campus in Cambridge, England. It was a great opportunity to take a peek inside the company that enables pretty much every piece of mobile hardware we use on a daily basis, and it was interesting to see where ARM is headed in terms of revenue growth and new markets. In case you're not familiar with ARM, their the company behind the architecture - the blueprint for a processor - as well as CPU and GPU core designs that end up in the likes of Qualcomm's Snapdragon and Samsung's Exynos processors. The fact that ARM operates a business model that licenses out their technology to partners has allowed the Android world in particular to explode over the past near-decade now. As a result of smartphones from all kinds of manufacturers using different processors all based on ARM's blueprints, we can have apps and games that run similarly on any device.
The smartphone and tablet world is a well-established one though, and the market is starting to show some serious signs of slowing down. ARM can continue producing reference designs for CPUs and GPUs for years to come, and considering the performance of the current crop of Cortex-A72 chips, it's almost getting to the point where we have all the performance we need. Where ARM sees a considerable growth in revenue is perhaps not surprising to fans of the tech world, the Internet of Things 'movement' has been around for a lot longer than many think, but it's starting to catch on in the mainstream, and while IoT seems an easy place to head, ARM isn't approaching things in quite the way you'd think. Beyond smart lightbulbs and connected gadgets, ARM is looking to a future where a connected world works smarter, not harder.
Speaking with Gary Atkinson, ARM's Director of Emerging Technologies he tells me that ARM isn't interested in small-scale consumer products, but rather more "disruptive technologies" that will lead to "hundreds of million units and above". Gary isn't interested in chasing small-scale applicator in high volume however, and says that he wants ARM's work in IoT to be "impactful" and "change how we live". As an example of doing just that, Gary says that one example could be to save water in agriculture by creating smart and connected fields that are more efficiently watered. The example given to me compared a field using sensors to monitor water intake and adjust only when needed rather than a big, hungry industrial solution that pumps out water indiscriminately. As our conversation continued, things started to turn to emerging markets such as India, who just happens to be the world's largest consumer of water, with China second. In India particularly, older, less efficient irrigation systems are used which leads to a loss of water before we've even thought about how these hot countries swallow up water more greedily. Moving from wet to dry, Gary tells me that sensor packages need not be expensive, saying that low-cost "tiny little sensors" could be deployed as a sort of "forest fire detection grid", accuracy isn't the main concern because all such a grid would be looking for is a "large delta change in temperature". These sensors could deployed in forests far away from towns and cities, and relay simple information such as "this is hot" or temperature increases that would definitely hint that a forest fire is on its way, allowing emergency services to respond quicker and potential save more lives.
All of this talk to use IoT for genuine uses seems something like Google, now Alphabet, would have come up with. When I put this to Gary, he says that this is "not a new idea" and refreshingly, neither Gary nor ARM is coy about the fact that deployment scenarios like these would lead to "millions of sensors" sold, leading back to royalty and licensing fees for ARM. The question is however, how will they do it?
The mbed Equation
ARM has been around since the 1990s, and those of us old enough to remember computers like the Acorn and BBC Micro (which has recently been revived as the micro:bit) can tell you that computers weren't also small or compact. For a long time now, ARM has specialised in teeny, tiny processors and how ARM is going to enable their partners and emerging businesses all over the world is by producing even smaller processors that require very little juice and last for a mind-bogglingly long time on a single charge.
mbed is ARM's answer to creating super-small and efficient processor and sensor packages to be deployed for pretty much whatever the user or business would want. Partnering with IBM, ARM's new platform leverages the power of the cloud to do all the processing and number crunching on powerful servers hosted all over the globe. mbedOS is lightweight, and runs no more than it needs to in order to relay information to the internet and give users and businesses the information they were after.
I was treated to a little demo of devices that were retrofitted with some off-the-shelf components running mbed and small ARM processors. My favorite example was of a Nespresso machine, a pod-powered espresso machine, which was rigged with a color sensor and small board that processed the information. A purple pod - the Arpeggio flavour - would register on a simple web app when inserted and the idea here is for businesses to register how many pods of each flavour they've used and to better keep an eye on inventory. There was a simple, and charming, use of NFC tags in a wine rack. An NFC tag would register as being removed when taken out of the rack, and the restaurant would then know that they've used up one more bottle of that wine.
More than all of that however, the most encouraging takeaway was the insane battery life of these small devices and beacons. These beacons are used for very simple, yet handy uses, and they can run for months off of small button cell batteries. A smartwatch they had on display was running mbed as a test case, and it registered battery life in months, not days. Needless to say the G Watch R on my wrist felt embarrassed during this demo time. mbed is soon to launch in a Technology Preview 3.0 and this seems the perfect new beginning for IoT. Rather than a proprietary this and a closed-off that, ARM is happy to license to whoever wants to join the party, thus pushing the whole market forward. Battery life is measured in months, devices are built for the cloud and the platform is flexible enough for anyone to do what they want with it. We'll be covering more on mbed, including how to get started with your own projects over the next few weeks, so stay tuned!