The automotive industry is currently undergoing multiple fundamental paradigm shifts at the same time: Electrification, higher levels of automation, or monetization of after-sales driving data for customers (B2C) and third parties (B2B). Younger automotive companies have in many cases taken the chance of adopting principles from modern software engineering from other domains; to let their product life cycle management be consequentially driven by customer usage data from the field. However, established automotive OEMs are facing organizational challenges to transition towards software-defined products to ensure, and even accelerate, the capacity of innovation.

In the near future, we are expecting a tighter coupling between the vehicles in the field and their virtual companions, or digital twins, at the automotive OEMs’ data centers. This coupling allows for (1) predictive maintenance, (2) safe preparation of virtual callbacks, (3) a more realistic and accelerated feature validation before roll-out, and (4) diversification of product features to smaller niche markets, not only by region but also by customer preferences across regions.

These software-defined innovations and products require a fundamentally re-thought system architecture for the platforms in vehicles as we expect a higher degree of concentration of computational capabilities (central and high-performance computing nodes) in the vehicle’s system and network architecture. However, just having more powerful computing capabilities is only the necessary but not the sufficient precondition to ensure a competitive level of innovation for the value creation chains in the automotive industry. Exploiting such high-performance computing requires a flexible, highly accelerated, yet safe and secure software environment that is not only ready for the complex and demanding software stacks to come for higher degree of vehicle automation (SAE Levels 3, 4, and 5), but that also is native to the complete end-to-end process between the vehicle platforms and the virtual companions in the cloud. As such, vehicle computing nodes must not be understood as separate and isolated islands but as part of a fundamentally different way of working that focuses on service orientation and value creation that enables the automotive industry to scale.

While AUTOSAR enabled the automotive industry to standardize decentralized and deeply embedded systems for dedicated vehicle features, a next generation software environment enabled by open standards and open source is needed that is powered by cloud-native methodologies to plan, remotely monitor, and systematically analyze system performance, and safely roll out and roll back new vehicle features.