A Framework for Cloud-based Testing of Multi-variant Cyber-physical Systems

Cyber-physical systems and IoT applications such as robotic production cells usually offer manifold configuration possibilities, e.g., with respect to the number and position of robotic arms, grippers, and movements and handling actions. This makes it hard to test such systems, since there is a huge number of possible test cases, and the execution of each test case involves complex simulations of physical interactions. In this article, we describe a method and tool chain which allow to configure and execute test suites for multi-variant cyber-physical systems in the cloud. We use feature modelling for describing the variability, a keyword-based approach for the specification of tests, and lightweight containers for the deployment of testing tasks onto computational resources. Our results show both the feasibility of the approach, as well as pointing out directions for future improvements.