Mixed-criticality systems, where multiple systems with varying criticality-levels share a single hardware platform, require isolation between tasks with different criticality-levels. Isolation can be achieved with software-based solutions or can be enforced by a hardware level partitioning. An asymmetric multiprocessor architecture offers hardware-based isolation at the cost of underutilized hardware resources, and the inter-core communication mechanism is often a single point of failure in such architectures. In contrast, a partitioned uniprocessor offers efficient resource utilization at the cost of limited scalability. We propose a partitioned real-time asymmetric architecture (PaRTAA) specifically designed for mixed-criticality airborne systems, featuring robust partitioning within processing elements for establishing isolation between tasks with varying criticality. The granularity in the processing element offers efficient resource utilization where inter-dependent tasks share the same processing element for sequential execution while preserving isolation, and independent tasks simultaneously execute on different processing elements as per system requirements.
翻译:混合临界系统,即具有不同临界水平的多个系统共用一个单一的硬件平台,要求将不同临界水平的任务隔离开来。隔离可以用软件解决方案实现,也可以用硬件水平分割实施。不对称的多处理器结构以未充分利用的硬件资源为代价提供基于硬件的隔离,而核心通信机制往往是这种结构中单一的失败点。相比之下,分割的单处理器提供高效的资源利用,以有限的可缩放性为代价。我们建议专门为混合临界水平的空载系统设计一个分隔的实时不对称结构(PARTAA),其特点是在处理要素中进行稳健的分割,以建立不同临界度任务之间的隔离。处理元件中的颗粒性提供了高效的资源利用,因为相互依存的任务在连续执行时有相同的处理要素,同时保持隔离,独立的任务则按照系统的要求对不同的处理要素同时执行。