Timed model-based testing

Testing is one of the most natural, intuitive and widely used methods to check the quality of software. One of the emerging and promising techniques for test automation is model-based testing. In model based testing, a model of the desired behavior of the implementation under test (IUT) is the starting point for test generation. In addition, this model serves as the oracle for test result analysis. Large amounts of test cases can, in principle, be algorithmically and automatically generated from the model. Most model-based testing methods deal with black-box testing of functionality. This implies that the kind of properties being tested concern the functionality of the system. Functionality properties express whether the system correctly does what it should do in terms of correct responses to given stimuli, as opposed to, e.g., performance, usability, or reliability properties. In black-box testing, the specification is the starting point for testing. The specification prescribes what the IUT should do, and what it should not do, in terms of the behavior observable at its external interfaces. The IUT is seen as a black box without internal detail, as opposed to white-box testing, where the internal structure of the IUT, such as the program code, is the basis for testing. In this chapter we will consider black-box software testing of functionality properties. Model-based testing should be based on formal methods: methods which allow the precise, mathematical definition of the meaning of models, and of notions of correctness of implementations with respect to specification models. One of the formal theories for model-based testing uses labeled transition systems (LTS) with inputs and outputs as models, and a formal implementation relation called ioco for defining conformance between an IUT and a specification [117, 118]; see also Chapter 11. An