Historically, hardware and software development have been tightly coupled and, in many cases, driven by long hardware-dependent cycles.
Deploying software inside a vehicle has been a complex and cumbersome process for automakers because of hardware disparities across the various makes and models in their portfolio, each with potentially different development and deployment environments.
For example, some cars may have up to 150 electronic control units hailing from a collection of suppliers, with each responsible for a single feature. Solving these problems is obliging automakers to find ways to decouple software and hardware development, while ensuring that software functions can be deployed seamlessly to vehicle hardware.
In fact, Arm’s latest announcement, the SOAFEE reference implementation, is the result of automakers, system integrators, semiconductor, software, and cloud technology leaders coming together to define a new open-standards-based architecture for the software-defined vehicle. Collectively, as defined by a Special Interest Group (SIG) of these leaders, the technology will be open-source software aimed at allowing broad prototyping, workload exploration and early development.
Software-defined vehicles decouple hardware and software while helping automakers and suppliers build developer-friendly platforms to access data, tools and machine-learning capabilities to develop, test and deploy machine-learning models in production within the vehicle reliably and efficiently. The result will be cost savings, higher developer efficiency and faster innovation.
A holistic approach to software-defined vehicles also enables cloud-native capabilities to be deployed on new consolidated, high-performance computing in-vehicle domain controllers. Vehicle functions run on top of these units and can be defined by software that is tested, simulated and delivered from the cloud that facilitates collaboration both inside the company and the value chain, thereby accelerating innovation.