Source: Arm Author: Arm
    We are ushering in a new automotive era, the era of software-defined vehicles (SDVS). According to Counterpoint Research, by the end of 2026, more than 1 million cars equipped with L3 level ADAS (Advanced Driver Assistance System) are expected to be on China's roads. Predictably, as the demand for high-performance computing and more software grows, so does the amount of computing power needed in cars. Given that future AI-enabled software-defined cars will contain up to a billion lines of code, coupled with significantly increased connectivity, the safety challenge is becoming even more acute. In order to avoid the serious impact of security breaches, the automotive industry has begun to take action to build deep security defenses throughout the SDV.
    Unlike the home environment, vehicles have unique "mixed-criticality" requirements that guarantee safety and real-time functionality while simultaneously supporting a more personalized user experience and a higher level of autonomy. This is what makes the car different from other scenarios, and also makes a lot of work carried out by Arm in the automotive field very targeted.
    In terms of addressing safety challenges, Arm this year brought a new set of automotive technologies designed to meet the higher performance, functional safety and information security requirements of AI-enabled SDVS. This includes a new range of Automotive Enhancements (AE) IP processors based on the Armv9 architecture, with the latest Arm information security features at the heart of the design.

afety considerations for three key SDV applications
    In addition to the general safety challenges facing the industry, there are specific automotive use cases that have significant safety implications for SDVS. These include digital cockpit/in-vehicle infotainment systems (IVI), Advanced Driver assistance Systems (ADAS) and autonomous driving (AD), as well as microcontrollers (MCUS) and regional architectures.

Smart Cockpit /IVI
    The increasing integration of smart cabins and IVIs in SDVS adds complexity to creating and subsequently managing these systems. As the connectivity capabilities and requirements in these systems continue to increase, so will the attack surface.
    The attack surface of the smart cabin is large because it covers multiple aspects, including cloud connectivity, connectivity to personal devices such as smartphones, USB plug-ins, and the ability to download apps. Hackers also have different motivations for breaking into smart cabins, which contain personal data that could be highly valuable to hackers, such as payment information. For IVI, the main security risk lies in providing a gateway to the rest of the vehicle that can be used to steal or control the vehicle. That leaves the door open for extortion or denial of service attacks.
    The smart cabin and IVI also meet advanced functional safety requirements, namely compliance with ISO 26262 and ASIL Class B security use case standards, which also require additional information security measures. Both systems integrate both safe and non-safe multiple displays, as well as a single physical display that combines functional safety requirements with other relevant information required by passengers and drivers. This creates a hybrid critical security environment that needs to be managed effectively from an information security perspective.

ADAS and AD
    The integration of ADAS increases the data volume and value of on-board assets. These include sensor and actuator data, AI models and algorithms for perception and target classification, graphics-intensive calculations such as 360-degree cameras, and a variety of mixed key considerations. The amount of software continues to increase, the potential attack surface continues to expand, and ADAS and AD directly affect vehicle control, so this may lead to further security threats.

MCU and regional architecture
    n the past, vulnerabilities in automotive MCUS were limited to attacks inside the car targeting specific automotive electronics, such as door mirrors. However, as the automotive industry accelerates its shift to more integrated and connected vehicle architectures, hackers can launch remote attacks on entire systems from outside the vehicle with increasingly connected components. This means that all MCUS need to have security measures in place, such as safe start, secure communication and verification support, in order to more effectively protect the vehicle. Even areas of the SDV that do not pose a significant immediate threat, such as automotive MCUS, need to be secured, as these areas could pose a weak link for hacking into other high-risk computing systems in the vehicle.

Arm is revolutionizing the automotive industry
Automotive Enhancements (AE) IP series improves performance and safety
     For nearly three decades, Arm has focused on providing information security-centric architecture capabilities that ensure the security of businesses, individuals and devices through an industry-leading technology ecosystem. We are working with our ecosystem partners to bring the latest Armv9 architecture security features, while also promoting continued collaboration on standardization and open source software to bring solid security to the entire automotive industry.
    Earlier this year, Arm announced a family of Automotive Enhancements (AE) IP designed specifically for automotive applications, designed to deliver high-performance, low-power, high-security computing power. It covers several core components including the Neoverse V3AE CPU, Cortex-A720AE processor, Cortex-A520AE processor, Cortex-R82AE processor and Mali-C720AE ISP to meet the full range of automotive computing needs.
    The new Arm AE IP incorporates key Armv9 defensive execution technology and architectural features to prevent attacks or malware. Techniques such as pointer validation (PAC), Branch object identification (BTI), and memory tag Extension (MTE) address the risk of growing lines of code by protecting the integrity of software control flow and reducing the impact of memory security breaches. This is critical for the automotive market because there is still a lot of legacy code written in non-memory-safe languages such as C that can be ported to future SDVS.
    In addition, Arm follows advanced product safety practices and standards, such as ISO/SAE 21434, to ensure that security risks are rigorously managed throughout the design, development and post-development phases of all products. Arm provides automotive partners with a range of supporting safety materials to make it easier to integrate Arm's off-the-shelf components into ISO/SAE 21434 compliant designs.
    Security is not achieved at the hardware level alone. With framework and API solutions, Arm is helping the software ecosystem deploy these architectural capabilities for superior functionality.
    Arm is actively involved in the creation of standard security apis and continues to contribute to them, such as the PSA Certified Encryption API, which builds a bridge between firmware developers and hardware vendors. In this way, Arm enables developers to focus on firmware design without having to understand the proprietary hardware rules involved with each new integration. At the same time, for hardware vendors, standard apis lower the barrier to entry, allowing them to focus on the value created by business differentiation.

Systematic design to accelerate the realization of the SDV future
    In addition to security, Arm is thinking about the challenges facing SDVS at the system level, such as standardization and the go-to-market process, seamlessly integrating hardware, software and ecology into a comprehensive solution that has excellent scalability, performance and energy efficiency, and can accelerate the product to market, and thus accelerate the SDV landing process. To this end, Arm introduced the Computing Subsystem (CSS), just like the Neoverse CSS and terminal CSS, Arm also plans to introduce the automotive CSS, which will pre-integrate and verify the configuration of its AE IP, and optimize the performance, power consumption and area on the advanced foundry process. The first Arm Automotive CSS is expected to be delivered in 2025.
    In addition, to help partners further achieve chip and software development and deployment, Arm also launched a new virtual prototype platform based on cutting-edge technology. With the virtual prototyping platform, Arm's automotive partners can evaluate IP through virtual prototyping without waiting for the physical chip to be ready. By reshaping the partner design process, the virtual prototyping platform accelerates the development and deployment of SDV chips and software, thereby shortening the time-to-market process.
    On the ecological side, to help the automotive and cloud computing communities build a common platform to facilitate collaboration and testing, and thus promote the SDV era, Arm also led the establishment of SOAFEE (Scalable Open Architecture for Embedded Edge) three years ago. SOAFEE is an industry-wide initiative in which Arm plays a leading role and promotes software standardization through a unified architectural framework that promotes interoperability and enhances ecosystem collaboration. SOAFEE's members have created a new ecosystem of software solutions to support the chip development and deployment process by enabling software consistency, which is critical for the Arm Automotive Computing Subsystem (CSS) to be launched in 2025.
    Since its establishment, SOAFEE has continued to grow and the number of members has exceeded 140, including chip suppliers, software providers, system integrators, cloud service providers, tier 1 suppliers, OEM manufacturers, etc. The Chinese community members include Geely, Lenovo, Zhongke Chuangda, Zhicong Technology, Zhida Chengyuan, Internet of Things Zhixing, Yingchi Technology, etc. Today, SOAFEE is moving to its next phase, "Soafee.next," which marks the start of SOAFEE's implementation phase.
    As a cornerstone of the automotive industry into the future, Arm has always been at the forefront of SDV technology innovation. With comprehensive safety features, an open, standardized platform and a strong ecological collaboration, Arm will continue to drive safety and reliability in the automotive industry, providing solid support for future AI-enabled SDVS.