In case you hadn’t noticed, today’s automobiles are becoming increasingly digital devices. In fact, one might even argue that the next major computing device category is the car. As enticing as that vision may be, however, it has also run into several roadblocks. Automakers are not computer makers, and the gap between the skill sets required and the skill sets available is real.
Huge, touch-enabled digital dashboards entice us into interacting with cars in ways that feel very similar to how we work with tablets, smartphones, and other digital devices. New features uncovered through those displays provide access to information about our automobiles and the world around them that makes us realize how connected cars actually are these days.
And let’s not forget that we’re consuming the same kinds of digital services within these modern cars: streaming music, video and other entertainment content, messaging, and various flavors of search, that we consume on our other digital devices.
Toss in the possibility of things like neural network-powered autonomous driving features, and well, yeah, it really is hard not to see our cars as the next major inflection in personal computing.
As enticing as that vision may be, however, it has also run into several very real roadblocks. Most notably, automakers are not computer makers, and the gap between the skill sets required and the skill sets available is very real.
To put it bluntly, it’s hard to turn a company that has focused decades of effort on improving mechanical engines and bending sheet metal in clever ways into one that knows how to create and leverage cloud-native software and a business model built around add-on services. Conversely, it’s a lot harder to achieve the gritty manufacturing and practical expertise of a carmaker than a lot of tech-focused companies initially thought (or were willing to admit). Witness the complete lack of actual vehicles from the many major tech powerhouses that have been rumored to have been working on automotive projects for many years now.
While it would be easy, and probably a bit fun, to speculate as to why we are where we are when it comes to the blending of the automotive and tech worlds, the simple truth is that finding the right mix of capabilities and industry knowledge is just plain difficult. It’s also been made significantly harder by the fact that virtually every carmaker has had to start its digital car efforts from scratch. There hasn’t been a common set of standards, frameworks and methodologies, and shared software components that would allow the automotive industry to advance together as a group.
The effort is designed to bring the open-source, cloud-native methodologies first enabled by major cloud computing providers like Amazon (AWS is part of the consortium) to the automotive industry. Of course, as its more open-ended name suggests, the group also hopes to eventually bring these types of standardized tools and capabilities to other categories of edge computing devices in the future. For now, however, the focus is clearly on the software-defined car.
Given Arm’s critical (though often overlooked) role in the automotive industry, it makes perfect sense for the company to have driven the creation of these standards and this consortium.
Arm’s complete IP range of simple M Series microcontrollers, R Series real-time computing devices, and powerful A Series SoCs are used throughout probably every car being made today via its chip-design and manufacturing partners — including companies like Renesas, NXP, Qualcomm, Marvell, Cypress Semiconductor, ST Micro, and many more.
The company is also uniquely positioned to take advantage of the knowledge and expertise it has gained in creating chip IP designs for the embedded edge via its Project Cassini, as well as the software and security requirements for IoT via SystemReady. Finally, let’s not forget its growing influence on cloud computing infrastructure via its Neoverse server and datacenter architecture designs.
Arguably, it is from this combination of experience that Arm recognized the opportunity to extend its learnings into the automotive space. Plus, while many other companies have been discussing software-defined, well, everything, as the future of many product categories — including automotive — Arm has the added benefit of creating designs that meet the challenging functional safety requirements of the auto industry, such as ISO 26262. It’s the combination of automotive-grade safety along with cloud-native capabilities and real-time operations that make the SOAFEE effort such an intriguing and compelling one.
Clearly, other key players in both the automotive and tech industries have been inspired enough by this vision to join the effort. In addition to many of its traditional silicon partners, Arm has also lined up CARIAD, a Tier One supplier that’s part of Volkswagen Group, Continental (another large tier one supplier to several major car OEMs), Red Hat, and the previously mentioned AWS, among many others, to join the SOAFEE Special Interest Group.
In addition, Arm has worked with chipmaker Ampere and IoT device maker Adlink to create two hardware development boards designed to be used by automotive supply chain players and carmakers to create and test their software applications. The AVA Developer platform features a 32-core Neoverse-based Altra SoC from Ampere and the AVA-AP1 offers an 80-core Altra SoC that’s intended for in-vehicle testing and prototyping.
On the software side, Arm has built several tools based on modern software technologies like containers, as well as CI/CD (continuous integration/continuous delivery) methodologies. They’re designed to allow software developers who have experience writing cloud-based applications to do work in the automotive space, without having to learn all the idiosyncrasies of automotive demands and requirements. In sum, they’re designed to allow companies to develop and run their automotive applications in the cloud and then deploy them in the vehicle.
Plus, by leveraging open-source principles, the consortium helps to build a library of software components that can be used by multiple companies, thereby dramatically increasing the rate of software advancements in the automotive space. This, in turn, should eventually lead to things like software-based feature/capability upgrades and a whole new set of business model and customer interaction possibilities that car makers have never been able to offer before.
Of course, talking about standards and industry consortia is a lot easier than making tangible, measurable impacts based on those ideas. Given Arm’s unique position in the automotive supply chain, its related experience in other categories, and the impressive set of partners with which this new industry standard is being launched, however, it certainly seems like SOAFEE is off to a solid start.