From requirements monitoring to diagnosis support in system of systems

Abstract

Context and motivation: Complex industrial software systems are often systems of systems (SoS) whose behavior only fully emerges during operation. Techniques such as requirements monitoring thus have to be used to observe such systems at runtime to detect deviations from their requirements. Question/problem: However, the focus of existing monitoring approaches is mainly on detecting violations of expected behavior, while support for subsequent diagnosis of violations is rather limited and often even neglected. Diagnosis is particularly challenging in SoS, which are characterized by complex heterogeneous architectures and a slew of different development and testing tools. Principal ideas/results: In this research preview paper we discuss the required capabilities for diagnosis support in SoS and outline a tool-supported framework based on a runtime artifact model and pre-defined diagnosis actions. Contribution: We describe our ongoing development of the framework and tools for supporting diagnosis in SoS and provide a research agenda.