Vehicle E/E-Architecture: Reduce to the Max: Page 3 of 6

November 27, 2019 //By Karsten Michels, Continental
Vehicle E/E-Architecture: Reduce to the Max
Over the decades the electronic architecture of cars has developed into a highly complex network with 70 or more electronic control units. Driven by the global automotive mega trends of “connected, automated, electrified, and shared” there is a fundamental change going on now towards a centralized server-based architecture. What will it look like in the future?

Enter the Automotive Server

This approach to a server-based architecture is a major transition because it is a radical change for the industry. The benefits of a server architecture, however, are so great that the shift has begun. By centralizing computing power and memory in a few in-vehicle servers, it becomes possible to provide sufficient and flexible capacity for applications. The data interchange within a server and between them is smooth and fast as it will be based on powerful communication technology such as GB Automotive Ethernet, for instance. And the underlying strict separation of hardware and software means that (new) functions can be almost freely allocated to the right place.

Different operating systems such as Adaptive Autosar, Android, Java VM, and Automotive-grade Linux can be run safely on the same hardware through virtualization with hypervisor technology and through applying lightweight container solutions. ASIL-rated functions, which require certain measures to ensure the higher level of availability and reliability, can be run on a separate micro-controller, called companion controller.

The in-vehicle server will offer several times the computing power of conventional automotive μ-controller systems which typically have between 500 and 3000 Dhrystone MIPS (= Million Instructions Per Second). The computing power of the initial Continental in-vehicle server generation is provided by a high-performance system-on-chip (SoC). Currently this groundbreaking product concept is being expanded into the Body & Security High Performance Computing platform (B&S HPC). Looking into the future, the Continental in-vehicle server is already designed for the extremely demanding durability requirements of a battery electric vehicle (BEV): In contrast to a car with a combustion engine, parts of the electrical system of a BEV are also active during the charging periods, which does include at least one of the servers in an electric vehicle. Hence this server has several times the operating hours of a server in a conventional car.

The end-of-line is no longer the end of the line

If a certain buffer of computing power is designed-in to an in-vehicle server, it will be possible to add new functions to a vehicle on the road. Probably one of the most important single advantage of a server architecture is its updateability over-the-air (OTA). One of the in-vehicle servers in the car will act as central gateway to all incoming and outgoing data and it will also provide core security features such as intrusion detection and verification of software certificates.

This allows to transmit and install software patches and security updates in a safe and well-coordinated manner. Providing software services to functions in the car will thus become possible as well. Benefits such as predictive maintenance will help to increase vehicle availability and to make necessary appointments at a work shop much less of a nuisance. Distributed computing will also become a reality, e.g. to facilitate the use of artificial intelligence (AI) by using the enormous computing power in the cloud.

Vous êtes certain ?

Si vous désactivez les cookies, vous ne pouvez plus naviguer sur le site.

Vous allez être rediriger vers Google.