At any level of computer networks, interoperability testing generally deals with several components that communicate while trying to provide a designated service. When a component remains silent, the assigned testing verdict is generally Fail, assuming that its behavior is non-conformant. Sometimes, this silence may be anticipated given the component's specifications. In these cases, the fail verdict is not unsatisfactory. In this paper, we show that "quiescence management" improves interoperability testing. Based on formal definitions of interoperability testing, we introduce new definitions that take into account the possible quiescence of components under test. Through several examples and scenarios, we show that these new definitions detect non-interoperability cases with higher precision. Moreover, these new definitions more clearly distinguish specification-driven quiescences from others, leading to unbiased interoperability tests with accurate verdicts.
展开▼
