Test Case Generation for Non-functional and Functional Testing of Services

Testing services offered at runtime is gaining more and more importance. This arises because the number of applications in which services don´t know each other a priori grows and occurring faults can cause high costs. The approach proposed here uses testing to assure QoS properties like e.g. real-time features and functionality of services. A challenge of testing is not to overload the service providing systems. Furthermore tests also have to adapt automatically to the currently offered services as well as to the usage of these services. We present an approach for testing services, which fulfills these challenges and is based on a genetic algorithm. To rate a functional result, the majority vote is used.