CAST-32A strategy for MCP certification builds upon the key terms of robust resource partitioning, robust timing partitioning, and robust partitioning. In this work, we provide technical insight on the interpretation of these terms by building on derivative terms such as resource capacity, capacity allocation, and capacity verification and violation. We discuss on the scope of applicability of these terms from the low-level hardware components up to the complete MCP platform. We also dig into the complexities in implementing robust resource and time partitioning, and show that those concepts must be applied bottom-up from hardware-shared resources up to the whole platform. We also show that, while robust resource partitioning is the preferred choice, it can only be realistically applied to some resources and, in many cases, robust time partitioning is the only choice. These considerations help to homogenize these terms, so far indistinctly used by software, hardware, and certification experts, as a building block to derive satisfactory approaches and tools to achieve CAST-32A objectives.
Type of publication
Publication in Conference Proceedings/Workshop
Year of publication
Place of publication
Automotive Technical Papers