How to address safety and security for software-defined vehicles

A New Approach Enables Different Architecture

As with any new advancements, vehicle technology has evolved over time. When technology was first introduced into vehicles, it was a series of domains and controllers. The next evolution moved to zonal compute platforms with multiple control centers or brains. But to realize software-defined vehicles that can interact with other entities inside and outside the vehicle, the next iteration must be able to support seamless connectivity and interoperability with other systems.

The most effective and pragmatic way of achieving this is the hybrid edge cloud computing model, in which all computing nodes within and outside the vehicle can act as cloud servers and connect with other systems at the API and microservice level. This approach futureproofs the automotive industry as demand for interactions with other entities and systems increases.

For automotive software engineering, this means the adoption of cloud and edge cloud technologies. OEMs will need to embrace cloud-native engineering principles with microservices, containers, zero trust and automation wherever possible. These modern technologies will allow vehicle manufacturers to take advantage of the coding techniques that SaaS software engineers use to improve agility. It will help automotive engineering deliver features faster, more efficiently and with significantly lower costs.

This transformation is possible by separating the critical from non-critical features. Critical applications must be on safe, proven platforms. By isolating the security and safety features, it is easier to maintain, diagnose and debug these features. This, in turn, streamlines certification processes by ensuring these features are not integrated with non-essential functions such as infotainment.

Alternatively, non-critical apps can go on containers and utilize the benefits of microservice architecture. This strategy means that agile teams can add and update features quickly with no negative impact on security or safety. And these non-critical features can be built on open-source platforms. For example, Linux is a good choice for non-critical customer-centric features and could promote faster innovation.