Towards a Model-driven Testing Approach for Microservice Architectures in the Automotive Domain

Connectivity and digitization have turned vehicles into smart devices which continuously increasing their functionality by interacting with their environment and sharing multi-modal mass data among people, businesses, service providers, or OEMs within the context of the Internet of Things. While microservices features scalability and flexibility , they are especially suitable for realizing mobility services that are deployed in cloud infrastructures. However, as connected vehicle scenarios have to consider environmental conditions, such as a changing connectivity, the according cloud-based mobility services must be extensively tested at different stages of the development process to get continuous feedback on how well-designed the software architecture and the service itself is. Nevertheless , testing requires mass data from real world scenarios to ensure both the proper functionality of the service and that the architecture is robust and scales with a varying number of vehicles. This paper addresses this problem by proposing a model-driven test framework for the virtual validation of cloud-based mobility services. Therefore, different simulators will be coupled within a co-simulation environment to generate massive amounts of vehicle-specific data and simulate the environmental conditions in which vehicles operate. This allows a scenario-driven validation if a service react flexibly (scalability), reliably (robustness), and with the intended behavior to service requests from vehicles.