On SMC-Based Dependability Analysis in LoLiPoP-IoT Project

Many systems require certain level of dependability to fulfill their purpose in predefined conditions. To check whether such a requirement can be met, the designer of a system must use proper means to assess dependability qualitatively or quantitatively, whereas this paper focuses on the latter assessment manner. The first problem with the assessment is that we cannot judge it except by evaluating its subattributes such as reliability, availability or maintainability. The second problem relates to the assessment itself - ideally, assessment builds on an analytical solution; however, if it does not exist, its presumptions are violated etc., an alternative approach must take place. This paper presents our alternative, simulation based approach with a special attention paid to reliability and maintainability; it builds on stochastic timed automata, an instrument able to model a wide class of systems/conditions of one's interest. In our approach, the assessment process takes the advantage of the statistical model checking technique, powerful enough to quantify dependability attributes in realistic situations and with a predefined degree of uncertainty. Finally, the paper evaluates our approach, outlines our research perspectives and gives a conclusion.