The SAMBA Approach for Self-Adaptive Model-Based Online Testing of Services Orchestrations

Service Oriented Architecture (SOA) is a popular design pattern that allows building applications composed of loosely-coupled and autonomous services. Such services may evolve and change at runtime, often outside the control of the owner of the application. Consequently, typical validation approaches, like offline testing performed before services deployment, are necessary but not sufficient: offline testing cannot assure the correct behavior of the SOA during its execution. To cope with the evolution of services and their orchestrations, in this paper we present a Self-Adaptive Model-BAsed online testing framework called SAMBA. SAMBA aims to assess the proper behavior of a SOA during its lifecycle executing model-based online testing at runtime, under the coordination of a MAPE-K control loop. SAMBA is assessed in a case study, where its detection capability are proved through functional, mutation and fault injection tests.