| Abstract: |
Hard real-time systems research defines different approaches to verify the performance in term of schedulability of systems, depending on its characteristics (e.g. the scheduling policy, multi or mono processor, periodic/sporadic events, precedence relation between tasks, network communication between resources, mutual exclusion resources, etc.). Whatever the approach selected, the first step consists in capturing the temporal properties of the system. UML-MARTE standard [1] allows to express those information, but its expressiveness is often considered too difficult for end users, and no UML-MARTE model editor has been widely adopted in the industry mainly because of these difficulties and the fact that the output could not be used directly by verification techniques like schedulability analysis or simulations. Depending on the real-time system characteristics, some methods and tools will be able to analyse the systems while others will not. Choosing the best tool to analyse one model is a task that requires a lot of strong expertize in real-time systems and a deep understanding of the tools capabilities. However, it is still time and effort consuming. The Time4Sys innovative approach, is not just to help users to design their system but also to select the tools that are applicable relevant for their performance verification, and potentially, and if needed to perform a conservative transformation in order to make the model analysable by tools that otherwise would not have been able to cope was not meant to be used for with this specific case. At the moment, Time4Sys has been successfully connected to MAST [2], Romeo [3] and Tempo Verifier (Thales internal proprietary tool). |