ETMCC: Model Checking Performability Properties of Markov Chains

Model Checking Markov Chains Markov chains are widely used for modelling the behaviour of computer and communication systems. A common goal of modelling is the estimation of performance and dependability measures of the system under study, such as throughput, buffer occupancy or mean time to failure. With model checking, the angle is somewhat different: Instead of conducting an analysis whose aim it is to derive measures, one first specifies requirements (with the help of a temporal logic) which are then checked automatically. Model checking of stochastic systems has developed into an active area of research. The logic PCTL, an extension of CTL, and the associated model checking algorithms allow one to specify and verify properties of DTMCs and Markov decision processes. In the continuous time setting, the logic CSL [1, 2] has been developed for model checking properties of CTMCs. While PCTL and CSL are based on the labelling of states with elementary properties (atomic propositions), the logic aCSL [3] is based on action-labelled transitions, which is advantageous if used in the context of modelling formalisms such as stochastic process algebra where the basic unit of behaviour is that of a named action. The Erlangen-Twente Markov chain checker (ETMCC) is a software tool which supports the automatic checking of performability properties given as CSL or aCSL formulas. The model to be checked is a finite CTMC labelled with atomic propositions and/or transition names.