Distributed system contract monitoring

Runtime verification of distributed systems poses various challenges. A pivotal challenge is the choice of how to distribute the monitors themselves across the system. On one hand, centralised monitoring may result in increased communication overhead and information exposure across locations, while, on the other hand, systems with dynamic topologies and properties are difficult to address using static monitor choreographies. In this paper we present mDPɪ, a location-aware π-calculus extension for reasoning about the distributed monitoring scenario. We also define numerous monitoring strategies for a regular expression-based logic, including a novel approach in which monitors migrate to ensure local monitoring. Finally, we present a number of results which emerge from this formalism, justifying our approach.