Automotive manufacturers (OEMs) must shift their mindset and be willing to embrace open standards in order to make the software-defined vehicle a reality.
Guests:
Suraj Gajendra, Vice President of Products and Solutions for the Automotive Business Unit at Arm
Chuck Alpert, Automotive R&D Expert at Cadence
Steve Spadoni, Application Manager for Regional Controllers and Power Distribution at Infineon
Rebeca Delgado, Chief Technology Officer and Principal AI Engineer for the Automotive Business at Intel
Cyril Clocher, Senior Director of High-Performance Computing for the Automotive Product Line at RenesasSiemens EDA Vice President of Mixed-Signal and Virtual Systems David Fritz
Synopsys Senior Director of System Design Group Marc Serughetti
An extremely important aspect of software-defined vehicle (SDV) architecture is safety and security. How will this be integrated with SDV?
Serughetti: Safety and security are two key elements. Safety is self-evident. We're talking about people's lives. How do you introduce safety into the vehicle? How do you verify it and verify it during the process? How do you make the hardware achieve this? On the other hand, there is security. When you think about SDV, it also means that the architecture will be massively integrated around the ECUs and the chips used. This directly affects safety because now you have more software, you have more attack points targeting this software. You have more security concerns. These aspects are absolutely crucial in the development of SDV. I use the word "development" because this is also an important point that needs attention in the context of SDV. You have the development part, which happens before the vehicle leaves the factory. But this is only part of it. As the vehicle is maintained, we begin to see changes in the software side, which is a shift in mindset. How do you start to separate software development from specific standard operating procedures on the OEM side? These are combined together - safety and security - are very important parts of the SDV development and deployment process.
Spadoni: In the implementation of SDV, the challenge is engraved in the software side and the company DNA trying to achieve this goal. These original equipment manufacturers (OEMs) are centered on mechanics, and also centered on the vehicle. They think in terms of vehicle milestones, not in terms of software development. I see some of them starting to decouple the two, so their software milestones are separate from the vehicle milestones. Their prototypes are starting to become more like mobile phones, such as the iPhone, but this is just the beginning, and it is not universal. Everyone is at a different stage, which is interesting, but it also causes a lot of delays. Many projects are delayed due to software development.
Delgado: At the recent SAE World Congress, a discussant said that as software complexity increases, the number of vehicle recalls is also increasing. I completely agree that in the automotive use case, safety and security are required. We can slightly distinguish between hybrid critical environments and very tight deterministic workloads. An element of software definition is to ensure that the correct workload runs in the correct slot. Then, there is also the high-performance computing field to provide the necessary support for the user experience. In this centralized evolution, everyone is talking about software-defined. Everyone is talking about central computing, safety, and security. The reality is that, given where we are today, there is a time variable between the level of centralization and the level the industry needs to reach - it may be slow but certain, or it may be painful. However, as part of the requirements for software definition, some of the technologies surrounding virtualization must be enforced at the hardware level to ensure partitioning, to ensure that hybrid criticality is retained. When you can trust how the operation will occur, you can focus on efficiency, which is crucial for this type of system.
Alpert: One thing I haven't heard yet is the word "chiplet," which surprises me. Without chiplets, SDV (Software Defined Vehicle) cannot be realized. The reason is that it allows people to achieve plug and play. I want a low-end chiplet for low-end vehicles, or a high-end chiplet for high-end vehicles. This also allows different suppliers to compete. But the key is, you must be able to achieve plug and play. You must have universal interfaces and standards like UCIe, which will define that you can really achieve plug and play, and everything will work properly. I thought about the whole concept of chiplets, which is to decompose the problem. Many years ago, we used another term - chip. Then we said, "Let's do SoC. We can combine them all." Now we are basically moving in the other direction. Thanks to the advancement of packaging and 3D technology, you can now improve the interface between chiplets and return to this instant decomposition model. The second thing to mention about software-defined architecture is that no one here wants to see car manufacturers pay for a large amount of copper wire. We would rather they spend money on silicon rather than copper. The zonal architecture allows you to save a lot of wiring. In addition, copper is really heavy. The last area is safety. This is an interesting issue because we want to reduce costs, and for those at the ASIL-D level (that is, the ISO 26262 standard that everyone is applying), the simplest way to achieve it is redundancy. Not one chip, but two chips. But two chips or two cores are very expensive. It doubles the area. The key is how to use tools and technology to achieve ASIL-D with less area, power consumption, and cost savings. In our functional safety solutions, we do indeed look forward to providing a cheaper way for companies building silicon to achieve this goal.
In the short term, what actions need to be taken to promote the development of SDV?Fritz: Previously, the topic of cultural mindset was mentioned. These companies still have a business model that requires them to generate revenue through model years, and the skill set required for deploying SDV is very different from the traditional mechatronics approach. OEMs (Original Equipment Manufacturers) need to quickly reach a point where they say: "We have two paths that will converge at some point." Before that, these two paths are distinctly different. Their organizational structure, operational procedures, and so on, are all completely different. Otherwise, it just won't happen. If they proceed incrementally, there will be an increasing number of recalls because they are trying to outpace the capabilities of the hardware or software. Therefore, they need to adopt this approach, and we see some companies are considering this.
Delgado: The most pressing goal at the moment is to truly collaborate with the right ecosystem partners to provide solutions for OEMs so that they can develop in a diverse manner. This is crucial. It has always been Intel's approach. We help other industries to participate in an open way, providing open computing APIs, and then scalable to the chiplet products mentioned earlier. Intel was the first company to announce an open platform, where the ecosystem can bring chiplets before all the standard definitions. This collaboration is essential to truly provide the efficiency required for the technology that already exists in the market.
Gajendra: Standards and collaboration, this is really key. In the short term, we need to reach a consensus on this. Then, based on this, we need to build a higher level of infrastructure in all aspects.
Serughetti: I agree with the previous discussion about standard collaboration. Looking at everyone at this round table, we have all collaborated. We compete, but we also cooperate to serve the industry. This collaboration needs to continue and needs to serve the market. I agree with David's point—this is all well and good, but if the transformation on the other side does not happen, no one will accept it. These OEM companies need to transform. Ultimately, the transformation will happen because if they do not change, there will be others to drive the change. Look at what is happening in China. This will force the entire industry to transform, and then it will be survival mode. So yes, some of these companies must change their way of thinking, and some must change quickly.
Clocher: In this group, many are partners of Renesas Electronics, and if we cannot obtain the right intellectual property (IP) from Synopsys in time, we cannot build our products. At the same time, we are also changing the way we develop hard-core silicon company software. This is also part of the company's transformation. In the past, the company developed chips, developed SoCs (System on Chip), and is now committed to chiplets. For the entire industry, if we can all rent or purchase these products with a large sum of money, then it will be successful. It can be SDV or other valuable technologies. Through the software and hardware technology we have jointly developed, OEMs must be open-minded when we go to dealers to spend $20,000, $30,000, $40,000 to buy these cars, bringing value to us. This is not our job. But we need to work together to provide this technology, enabling OEMs to do better, simply by leveraging the future SDV platform and transforming it into new features and new user experiences.
Spadoni: From Infineon's perspective, the most urgent thing is the mentioned way of thinking. It's not just the way of thinking of OEMs or Tier, their organizational structure must also match it. The current situation is not the case, and in order to make SDV a reality faster, this must change. There are many competitors around the world doing this, and everyone has noticed this.
Alpert: The way of thinking is a good point, and this situation will happen. For companies that have not done this, they will not exist in 5 or 10 years. But it's also about speed. If the design requires a 5 to 7-year life cycle, how long will this take. The entire transformation and virtual platform must happen, which means that collaborative standards must be established. Companies that try to do everything on their own will not survive. They must be able to collaborate, but then figure out how they can differentiate. Whoever figures this out will win.
Comment