By Marques McCammon
Auto makers are under pressure to deliver more innovation at a faster pace than ever before, all the while maintaining high levels of safety. Typically, the introduction of new technologies means added complexity and greater cost. So, how can OEMs address evolving market needs and stay on time and budget? A key piece to the solution lies in taking back ownership of the software in its cars.
Traditionally, OEMs work with a diverse web of suppliers, with each developing their own components in a highly controlled and opaque manner. OEMs typically did not need to know or care what went into the component as long as the technology delivered on feature, complied with industry standards and stayed within costs.
In this post, we’ll be discussing the concept of Abstraction or the “A” in the Wind River ACRU model introduced in my last blog post. ACRU is a software strategy designed to help OEMs keep up with consumer expectations and upgrade legacy technology without constantly adding more cost, weight, power consumption, and complexity that eats away at profits.
Inside an automobile, you’ve got both hardware and software elements. Software is dynamic and fluid, and nowadays, it can be more readily changed. Hardware is a fixed element and it’s what will typically drive up the cost and investment structure to produce cars. Ideally, we want to get the maximum use and value out of the hardware. To that end, hardware longevity is key to making that possible; and one way to achieve this is to ABSTRACT the software from hardware, effectively minimizing the dependencies between the two. Abstraction can also maintain supply chain integrity.
When software is isolated away from the hardware and no longer directly dependent to it, OEMs can create an environment where competitive parties bid for the same piece of hardware business, whether it be at the engine control unit (ECU) level or the algorithm level, keeping total costs down. Having a flexible software architecture design allows OEMs to scale and evolve over time.
Borrowing a concept from the telecommunications and networking industry, Wind River takes the approach of multi-layered decoupling as a way to deliver new advantages. This starts with “silicon flexibility”— enabling integrated circuits from multiple suppliers to support the same software architecture.
The next layer is building a secure and reliable framework to act as the foundational core. By having a guiding framework that serves as a backbone for building multiple variants, it is possible to bring in new software algorithms and applications, whether from software vendors, from an OEM, or from a tier one supplier. Either way, it can be brought into the development environment faster, and in a way that’s more predictable, because there is a solid backbone that provides strength and integrity as the system evolves.
Wind River’s approach also abstracts the development process. In telecommunications, there often is an abstracted, virtualized compute environment that can move from one data center to the next seamlessly. In the automotive industry, a capability like that would be analogous to being able to shut down the assembly of car in one factory and then bring it up in another factory, within just a matter of days or weeks.
If auto OEMs can bring this abstracted approach and capability to the development side of the business, they can re-create the entire software stack, simulate their hardware, simulate their use cases, inputs and outputs, and dramatically reduce development time and cost— and even start debugging before deploying the software to new vehicles. This would translate to dramatically improved time-to-market, cost efficiencies, and scales.
In my next post, we’ll be discussing the “C” within our ACRU model.