ABSTRACT
Dependability analysis by Markov modeling is extremely straightforward, but the price paid for its simplicity is that it makes assumptions; the analyst needs to assess the applicability of these before applying the technique to a particular system. For any Markovian analysis to be valid, the following assumptions must be made about the system under analysis: When the system is in a particular state, the history of how it arrived in that state is irrelevant for determining what will happen next. Markov modeling can be quick and easy to perform and has the advantage of providing precise and exactly reproducible results. Its disadvantage is that it assumes that failure rates are constant, the interarrival times of failures being negatively exponentially distributed. To create a Markov model it is necessary that everything be expressed as a rate.
