More from ARM TechCon. Great show as always, high-energy and a reminder that systems and solutions are where it’s at. There was a very big focus on Internet of Things in all its many guises, from devices to detect whether a garbage container is full, to a child’s necklace to store immunization and other health data, to new ways to push entertainment directly to whatever screen we happen to have in hand. But for those who think the IoT is an easy path to fame and fortune, there were reminders that this domain still has plenty of challenges ahead.
Let’s start with venture investment. That sexy new wearable you think will blow the VCs away? Forget it. According to Eric Klein, a partner and VC at Lemnos Labs, the Consumer IoT market is “very saturated”. That aligns with a couple of frequently overlooked points in consumer economics. First, we only have so much disposable income. Even if a carrier is nominally prepared to underwrite the cost of a device, they are digging deep into your wallet for that plan. If they find a way to bundle an additional device into the plan, they’ll dig even deeper. Second, we all hit gadget fatigue at some point, even the uber-geeks. Life (for most of us) offers too many other distractions. Eric’s take-away – focus your creative energy on enterprise and commercial applications. Businesses don’t care about sexy; they pay to reduce costs and increase revenues.
Next up, security. An expert panel (Paul Kocher – Cryptography Research, Eduardo Montañez – Freescale and Zack Colby – ARM) discussed where they really think we’re at. ARM has done a lot of good work to build a turnkey security solution at the device level – TrustZone® with CryptoCell as the basis for a trusted platform, secure communication over mbed TLS, secure code compartments through mbed OS uVisor and secure lifecycle management through mbed ID, Config and Update. And GlobalPlatform will certify complete Trusted Execution Environments. All of which are good steps to get to to a unified approach to security in the IoT.
But another observation was that this only works if the people at the end of the value chain turn it on. There is evidence that at least some of those folks find that step too difficult, or too costly, or just not very important – until of course there’s a breach, at which point it will be your fault. So we have to deal not only with bad actors but also careless actors – in buying, installing and maintaining. Changing these behaviors won’t be as easy as standardizing software and hardware.
Last and very far from least, a few observations from Colt McAnlis’ keynote. He’s a developer advocate at Google and a very entertaining speaker. His main point is simply stated and sobering – that if we’re not careful, all the entrepreneurs who are busy developing zillions of new solutions are going to screw up the IoT. One example is obvious (after he presented it). You walk through a shopping mall; each business has B2C (biz-to-consumer) “things” to ping your phone as you walk by, alerting you to their great sale on jeans. Two problems here: the annoyance of constant pings and your battery draining thanks to that WiFi/BT communication. Then you click on the message/app which takes you to a website with a rich graphic (or even video) experience, requiring an LTE session, which drains your battery even more. Now amplify that to add automated parking stations, automated hotel doors, automated checkouts… Pretty soon you’ll need a backpack battery to keep your phone charged. And then someone (Colt’s daughter in the keynote) will develop an app to block all that stuff and B2C IoT (two acronyms in a row, a personal best) will die a quick death. That’s what will happen if we don’t coordinate to avoid annoying consumers.
Colt had more good stuff on not defaulting to using internet “classic” technologies and data on the IoT. An example he gave was JSON. Great method on the standard internet to pass around structured packages of data. All text, human readable, easily debugged, what’s not to like? The answer is a massively bloated format burning unnecessary power to communicate and pack/unpack on a cell phone. There are more compact formats like FlatBuffers which are much more efficient. Another “classic” use-model example is widespread use of pictures. We’re graphics junkies now but they’re very expensive to download. That’s not a problem if you can pick and choose but see above. In our website development kits, we need a new target “format”, along with laptop, tablet and phone. We need an “only marginally annoying IoT format” which will skip downloading ads, cute Flash videos (and while we’re at it, will download just the one page you need, in minimal text, not the home page, or a page where you have to scroll forever to get to what you need).
Colt made us hold up our phones at the end and swear we were not going to screw up IoT. If we want it to thrive, we’d better listen.