How automotive software development is changing, and what that means for standards, architectures, and long term flexibility

In this interview, Anders Hallgren, VP Product at SystemWeaver, shares his perspective on AUTOSAR, the shift toward service based communication, and why technology agnostic system design is becoming increasingly important for OEMs navigating rapid change.

What is happening with AUTOSAR?

From what we see through our participation and through our customers worldwide, the Classic Platform is still strong. It is designed for signal based communication, has been developed for many years and has a large ecosystem of suppliers providing basic technologies and stacks. It will continue to be an important part of the automotive industry for quite some time. However, there is also a clear paradigm shift. The industry needs more computational power in vehicles and is moving toward zonal architectures, which means shifting from signal based to service based communication. AUTOSAR calls this their Adaptive Platform. It has not gained traction in the way they likely hoped, and I believe this will create an entirely different situation for OEMs. The future may become more proprietary, or new standards and formats may enter the picture. The reason is that this technology sits much closer to how software and computers communicate in other industries. Because of this, companies that come from a background closer to computing, even personal computers and processors, now have space in the automotive market and are moving quickly. If we also consider the competitive landscape, China has been moving fast for many years. It is very impressive. Standards are good. We believe that. But standards also create inertia. The question is whether AUTOSAR will be fast enough to support what people want from the standard. If not, the industry might move in another direction. From a SystemWeaver perspective, we like AUTOSAR and make sure we stay compatible with the standard. At the same time, we remain agnostic to the underlying technologies. We give our customers a way to document and design their systems, regardless of whether their goal is AUTOSAR or something else. That is one of the strengths of our platform. It lets you abstract and stay agnostic, both between hardware and software and toward specific technologies. This is what we do, and have done for 25 years, so we know this space well.

What are we doing instead of AUTOSAR?

Instead of AUTOSAR alone, many new middle layer players are emerging. Some develop stacks for service based communication built on the ROS standard, which is quite different from how AUTOSAR usually works. This is the direction the industry is taking, and some organisations might even build their own solutions. What we do is offer technology agnostic solutions that support customers regardless of whether they use AUTOSAR, something else, or a mix of AUTOSAR and proprietary technology. This also includes solutions based on ROS, which is an open source standard for service based communication that has now entered automotive as part of this paradigm shift. We are well prepared to support our customers in whatever technical direction they take.

Looking ahead

I believe many things will happen. Change is already underway, driven from two directions. First, the macroeconomic climate is tough. This puts pressure on automotive companies to become more efficient, faster and more competitive with the same or fewer resources. It is becoming clear that maintaining control and visibility of information is essential. This avoids redundant work, which is a waste of time, and increases overall efficiency. In parallel, AI is advancing rapidly. A key enabler for AI is having organised, clear and accessible information. With proper information management, organisations can use these powerful tools and agents to become even more efficient. This is where the world is heading: larger scale work, involving more people, and mapping expertise to what needs to be done. We see this both in general development projects and in areas that used to be more specialized, such as functional safety and cybersecurity. I expect these domains to be carried out at a broader scale to increase efficiency and remove bottlenecks in processes.