Mike Muller, ARM’s chief technology officer, used his opening keynote at ARM TechCon to give away the formula for success in the Internet of Things (IoT): (performance x trust)/(energy x $).
Muller's one concern was whether trust was a multiplier or an exponential, but the emphasis on security and system integrity throughout the first day of the show seemed to indicate exponential. He was being conservative.
The root of all this mistrust around the IoT has to do with each of the >50 billion nodes that will be connecting to embedded systems by 2020 (according to Cisco). To date, adding security has been an expensive proposition, yet if these nodes aren’t secured they may expose those systems to hackers. This leads to insulin pumps, pacemakers, health monitoring information, and automobiles all falling under the control of remote ne’er-do-wells.
But security is hard: It’s hard to do and it’s hard to enforce, as 99% of a system may be bulletproof, but all a hacker has to do is move up a level to the least secure point in the system. That could be anywhere from the sensing node to the gateway to the server. The attack surfaces, as they say, are bountiful.
Mike Muller speaking at ARM TechCon.
That’s why Muller spent so much of his keynote time on security, even though he had a major core announcement: the new “ultra-high-efficiency” Cortex-A35. The Cortex-A35 extends the ARMv8-A 64-bit architecture into mobile and embedded markets, completely shattering the idea that 64-bitters are for servers only.
The core steps in as successor to the Cortex-A7, and at its 28nm node implementation is expected to consume less than 90 mW at 1 GHz and less than 6 mW at 100 MHz. Expect to see chips using the core by the end of 2016.
As exciting as the core may be, Muller emphasized another big announcement: its TrustZone technology would be extended to its new ARMv8-M architecture, also announced at the event. TrustZone is essentially a partitioning feature that lets you assign trusted and nontrusted zones, and keep them separate and isolated to prevent attacks to your system.
While adding anything to a microcontroller that may cost $0.01 is a big deal in terms of return on investment, that’s kind of the point Muller – and ARM – are making: Security matters, from the node to the system. You, as a designer or developer, can’t afford to skip it. Regardless of cost. In fact, it’s critical to the success of the IoT. Leading us back to the formula above.
But security was everywhere at the show, especially with automotive and the specter of autonomous vehicles being hacked. I attended a panel on advanced driver assistance systems (ADAS) and while the panelists discussed topics such as how difficult it is to make a good wireless connection – besides Bluetooth – in an automobile, they danced around the topic of security. So much so that the moderator had to bring it up and express surprise that he was forced to do so.
The reason may be that we know security is important--we also know we have do enforce it--but it’s kind of a buzz killer, especially given how pointless it can seem given how many holes are in any given system. Yes, we know we have to deal with it, we’ve been saying so on every automobile and IoT panel discussion over the past five years. Next topic, please?
The ADAS panel at ARM TechCon.
Get your automotive design idea noticed
That’s why I broke in and asked the panelists how to get into the automotive supply chain if you have a good idea for product or feature. With automobiles now a hotbed of innovation, instead of closed systems only accessible to Tier 1 suppliers, designers and tinkerers have the opportunity to provide interesting design ideas, especially within the lines of the IoT. How many of you have ideas that you’d like to see implemented in a car – and would you like to design those products yourself?
If the answer is yes, then the panelists (Scot Morrison of Mentor Graphics and Shrikant Acharya of Excelfore Corporation) had some sage advice: Despite the openness of automobiles to new ideas, manufacturers still prefer to work with Tier 1 suppliers.
So, if you have a good idea, find a Tier 2 supplier and see if you can interest them in working with you on the idea. If you prefer to go it alone, rather than wasting time reaching out to Toyota or GMC, make a consumer product that users would like to see in their automobile.
If the idea gets traction, that’s when automobile manufacturers will come to you.