If the development model of software-defined vehicles was only in its embryonic state and initial stage in the past, then with the advent of the data-driven and AI-driven era, software-defined vehicles have undoubtedly entered the deep water area, and the automotive software industry is rapidly advancing in the vigorous development.
The list of high-quality authors is well illustrated by a set of comparative data. At last week's ARM Tech Symposia annual technology conference in Shanghai, Dipti Vachani, senior vice president and general manager of ARM's automotive business unit, said that the software ** of a car is as high as 500 million lines today, and it will reach an incredible 1 billion lines in the near future, completely doubling. Just a decade ago, there were only 10 million lines of software in a car**.
At the same time, authoritative data also indicates that the market for software-intensive electronic systems will grow at a rate of about 15% per year until 2030. However, it is a rule that rapid growth is often accompanied by serious challenges, such as the increasing complexity and security of software, DevOps agile development issues, how to deliver quickly, how to ensure software portability and how to improve cost-effectiveness, all of which are headaches for car companies.
* The far-sighted ARM has long understood this. In 2021, led by ARM, with the participation of industry partners such as Continental, Volkswagen Cariad, Amazon Web Services, and Bosch Bosch, the Arm SOAFEE open architecture based on cloud-native technology is designed to solve a series of practical problems in automotive software development.
In the past two years, with the continuous deepening of software-defined vehicles, SOAFEE has also been vigorously developed, and has been favored by many manufacturers in the automotive field in the industry, and the ecological territory has been expanding.
Dipti Vachani stressed that Arm has transformed into a computing platform company, but it has also always adhered to the model of licensing IP, which will not change. Because of the growing demand for dedicated and vertical computing, Arm is also building a strong vertical computing platform for niche domains.
Currently, Arm has four broad categories of vertical computing platforms covering mobile, infrastructure, IoT, and automotive, each targeting different needs. Among them, ARM's comprehensive computing solutions are mainly for mobile devices, ARM's Neoverse platform is mainly for server and data center design, ARM Corstone is mainly for AI and machine Xi at the edge and the Internet of Things, and SOAFEE is mainly for the automotive industry and a new generation of software-defined vehicles.
According to Dipti Vachani, Arm's partners have always wanted a dedicated computing platform that can host or run their own workloads. It is in this demand that Arm set out to build dedicated computing power and develop solutions based on it, "The SOAFEE architecture is focused on solving automotive software development problems in the automotive field. As an open architecture with a strong ecosystem, SOAFEE has more than 100 members across the global value chain in just two years old. ”
Based on the advantages of cloud native, Soafee can solve a series of stubborn problems of traditional software development, such as software portability, accelerated software development and software upgradeability, and these three points happen to be the problems that Soafee is mainly committed to solving. So, what exactly is SOAFEE, what is its mission and core values, and why it is an ecosystem, we will give a detailed explanation below.
As you can see from the above description, SOAFEE is a large, collaborative community led by ARM, which means open source expert group, and the full name is Scalable Open Architecture for Embedded Edge. The evolution of the SOAFEE architecture and the development of the software-defined vehicle model are almost identical in rhythm.
Because in the process of software development, developers realize the importance of cloud-native technology and centralized computing platforms in automotive software framework design, sample reference, and testing, so the value of SOAFEE is becoming more and more prominent, because it happens to provide a system.
1. Scalable, standardized automotive software development architecture that supports cloud-native development of all workloads.
We can also think of Soafee as a collection of meritocracy. With their respective experience and strength in the field of smart cars, the members of the organization will participate in the definition of a new future-oriented cloud-native automotive software architecture and reference design, and will improve the software application ecosystem of SOAFEE based on their own rich smart car product solutions.
At the same time, all members of the organization can also absorb and learn from the experience and essence of other members, obtain open source sample references from the SOAFEE architecture, and lay the foundation for their own creation of open source solutions and ecosystems, which forms a virtuous circle of repeated and continuous upgrading.
We've been talking about software-defined cars all the time, so why is software defining cars?This is because the hardware of contemporary automobiles is gradually standardized and modularized, and the value and role of software in automobiles are increasing day by day. From intelligent network functions, autonomous driving and perception fusion and AI interaction, to big data operation and vehicle remote OTA capabilities and vehicle core systems, from system software, application software to intelligent algorithms, all need to be realized in a row.
This is the origin of what we said earlier, the increasing amount of **. This is a direct result of increasingly heavy workloads and increasing software complexity. Until the SOAFEE architecture came out, industry stakeholders found that this is a very good platform tool to promote automotive software development, cloud native, open source extensible, software portability, promote DevOps agile development and rapid deployment and non-stop delivery characteristics, which greatly improves the efficiency of automotive software delivery, but also ensures the safety and reliability of systems and functions.
Taking DevOps agile development as an example, Sensory, one of the members of the Soafee Special Interest Group (SIG), a world-renowned intelligent voice vendor, based on ARM Cortex-M7, relies on microservices and continuous delivery capabilities to develop and deploy its voice, vision and biometric AI technology and cloud-native applications, and the entire development process realizes the portability and rapid deployment of software between different systems without cross-platform compilation, cross-platform debugging and complex toolchains.
From the perspective of the core design goals of the SOAFEE architecture, it is a standards, developer resources, and professional processing platform designed for the security and real-time requirements of automotive applications. By defining a framework that supports cloud-native development of vehicle applications and functions and deployment of vehicle edge platforms, the SOAFEE architecture allows for the integration of different middleware and application software stacks. In addition, it seamlessly creates a seeding platform for cloud-to-automotive edge software development with cloud and endpoint consistency.
Its cloud-native features support cloud-native development for all workloads, including workloads with functional safety, time partitioning, spatial partitioning, and real-time requirements, all of which are important features of next-generation automotive embedded edge platforms.
According to Dipti Vachani, Arm's AE IP (Automotive Enhanced) integrates advanced safety features. In addition, Arm's IP is scalable, including CPU, MCU, MPU, GPU, and ISP. "Whether it's functionality, performance, or heterogeneous computing, there's this scalability from automotive MCUs to fully autonomous vehicles. ”
Thanks to the key value brought by the SOAFEE architecture, in just two years, SOAFEE has been valued and favored by more and more automobile manufacturers, semiconductor chip manufacturers, cloud technology leaders, software vendors, artificial intelligence algorithm and solution providers, and system integrators.
As mentioned earlier, SOAFEE has more than 100 members of the global industry value chain, and they are all heavyweights and representative vendors in the industry. Within the SOAFEE framework, they are willing to contribute and share, but also learn from others to absorb more nutrients and experience. And China's related manufacturers are also continuing to join this big family.
Joynext, a member of Joyson Electronics, one of China's top 100 software companies, officially joined SOAFEE as a voting member in 2022 to build a cloud-native digital foundation for automotive software and push in-vehicle basic software to a new level. Thundertech, a provider of intelligent operating system products and technologies, also joined SOAFEE in 2022 as a voting member of the working group, working with the working group members to continuously improve SOAFEE's reference software stack and jointly build an industry-leading computing architecture and development system.
It can be said that there will be more and more examples like this. Dipti Vachani also spoke highly of ARM's development and ecological construction in China. She said that there are currently nearly 400 Arm technology licensing partners in China, and the cumulative number of chips based on the ARM architecture shipped by Chinese partners has reached 30 billion to date. In addition, of the world's 15 million ARM developers, 4 million are in China.
It is precisely because of such a strong ecological foundation that SOAFEE will inevitably continue to enrich and grow, and eventually become a true standardized architecture for the automotive industry.
A significant trend is that in the future of automobiles, the value of software will be more prominent, the car will be reconstructed, the business model will also be subverted, and the development of various intelligent technologies and functions will be more defined by software, and even in the future, it will not be selling cars but selling software.
Developers at automotive manufacturers often complain that the software delivery process is often subject to various phases, such as the need for thorough testing before releasing alpha, beta, and production versions. With Soafee, its continuous integration and continuous delivery (CI CD) features allow the process to be fully automated. In addition, when used in tandem with DevOps, it allows project teams to collaborate on a larger scale, reduce risk from product release, and shorten the time from development to release.
And that's just one specific aspect of the value that Soafee brings. According to Dipti Vachani, 70% of today's machine Xi applications are already running on the ARM platform. In the future, there will be more examples of innovation and collaboration based on the ARM architecture, and the SOAFEE platform will continue to deepen to help software-defined vehicles move to a new stage.