Performance along with dependability analysis is a tremendous challenge in the design
or improvement of modern complex systems. Two different classes of solution methods
are generally used: analytic-numeric methods and simulation methods. As most of the
literature explains, the choice between them depends more on the analyst's background
than on the system itself. In this paper, we illustrate the advantages and drawbacks
of each method on real problems and compare the results. Finally we conclude the paper
providing some hints to choose a solution method depending on the model. We use SPNP,
a Petri net analysis package, and CSIM 19, as simulation package to model and evaluate
systems. Copyright 2007 ICST.
