ABSTRACT
While unit testing is supposed to guarantee the proper function of single units, integration testing (ITest) is intended to validate the communication and cooperation between different components. ITest is important because many events are caused by integration-related faults such as, for example, failures during money transfer, air-and spacecraft crashes, and many more that are not detectable during unit testing. This chapter introduces an approach to model-based integration testing. After a brief review of existing work (1) communication sequence graphs (CSG) are introduced for representing the communication between software components on a meta-level and (2) based on CSG and other introduced notions test coverage criteria are defined. A case study based on a robot-controlling application illustrates and validates the approach.
