Generalized metric based test selection and coverage measure for communication protocols
Jinsong Zhu, Son T. Vuong · 1997
This paper presents an important generalization of the metric based test selection and coverage measure, originally proposed in [14, 5]. Although the original method introduces a significant analytical solution to the problem of coverage and test selection for protocols, its applicability is limited to only the control part of protocols. We extend this method to handle a protocol behavior space where both the control sequences and the data valuations for event parameters are included. We prove that the important properties of total boundedness and completeness are preserved in the generalized metric space, thus ensuring the possible approximation of the specification (infinite number of execution sequences) with a finite test suite within arbitrary degree of precision. A generalized test selection algorithm and coverage measure are also discussed.