This paper describes PHYLOG, a framework intended to help certify the use of a multi-core in an aeronautical context. Specific guidelines for such systems have been published in a document, the MCP-CRI / CAST-32A, which provides a series of objectives to be fulfilled. To justify that an objective is indeed achieved, PHYLOG relies on structured graphical notations, recursively refining each objective by solving it using a strategy, itself having its own sub-objectives, until all that remains are evidences found either through the use of formal methods, as part of the design choices, or in external documentation. The PHYLOG framework includes such formal methods, providing the means to model the multi-core and to acquire further evidences through automatic analysis.

The CAST-32A provides some guidelines to help certify multi-core-based systems in the avionics domain. One major requirement is to compute all the potential interference and to provide adequate mitigation means. In this paper, we compare two approaches to identify the interference: the initiatortarget and the Phylog models. The latter is more compact and efficient, despite also covering all of the problematic conflictual situations.

No abstract

No abstract

In distributed systems, the order in which the messages are received by the processes is crucial to ensure the expected behavior. This paper presents a communication model which allows for restrictions on the deliveries of a channel depending on the availability of messages in other channels. This corresponds to prioritizing some channels over others. It relies on a framework able to verify if a given system satisfies a user defined LTL (Linear Temporal Logic) property with different priorities. We also propose to automatically infer the channel priorities so that the system does not infringe on this temporal property.