Don't lose control, stay up to date: Automated runtime quality engineering

Abstract

Modern industrial settings require great flexibility of systems. For example, in automated factories, smart cities or smart grids, systems are exposed to highly variable and dynamic environments: Not all events can be modeled at design time, system components may fail to operate as desired at runtime, and stakeholder requirements may change at high frequency. In order to handle these challenges, systems have to expose highly flexible behavior. Due to the complexity of application domains, deterministic specification and validation of system behavior and quality is no more feasible. Instead, new approaches for building flexible systems yield probabilistic behavior at runtime. While this defeats the purpose of design time testing and quality assessment, it is indeed possible to shift the standard design time engineering activities towards runtime. This enables monitoring and validating system behavior at runtime, incorporating information not available a priori and using it to keep the dynamic system under control at runtime.