Towards compositionality in execution time analysis: definition and challenges

For hard real-time systems, timeliness of operations has to be guaranteed. Static timing analysis is therefore employed to compute upper bounds on the execution times of a program. Analysis results at high precision are required to avoid over-provisioning of resources. For current processors, timing analysis is a complex task mainly due to interdependencies of the processors' features that affect the overall timing behaviour. To still obtain tight bounds, state-of-the-art approaches collect detailed information about these interdependencies by exploring the state space of the system as a whole. Modern systems, such as multi-core processors, introduce even more timing dependencies -- e.g. due to interference on shared resources between functionally independent programs running on different cores. This will eventually render the above, non-compositional, approach infeasible in terms of analysis runtime and memory consumption. Therefore, recent analysis approaches often assume a certain independence of system components -- referred to as timing compositionality. We aim at a formal definition of timing compositionality as it was previously only introduced informally. How to achieve timing compositionality in general is an unsolved question. We highlight challenges and summarise open problems that arise in the context of compositional analyses.