Hypervisor Technologies for Functional Safety in Cloud Computing

As pivotal sectors such as automotive electronics and rail transportation undergo digital transformation, there is a growing demand for enhanced performance, scalability, and functional safety in computing platforms. Many core servers, characterized by their high parallelism and substantial hardware resources, have emerged as optimal environments for hosting such applications. Nevertheless, current virtualization technologies encounter challenges related to performance degradation and resource contention when implemented in many-core architectures exhibiting Non-Uniform Memory Access (NUMA) properties. These challenges hinder the ability to satisfy the stringent requirements for real-time performance and safety. To tackle these issues, this paper introduces an innovative architecture for a virtualization environment (hypervisor) specifically designed for many-core processors. This architecture incorporates distributed deployment, resource isolation, and optimized communication mechanisms to effectively alleviate resource contention and minimize cross-die communication overhead while facilitating the efficient co-execution of applications with varying safety levels. The paper offers a comprehensive overview of the system structure and communication mechanisms associated with the proposed hypervisor.