In computer integrated manufacturing environments, dependability is a cruci
al attribute for the production management and control information system,
which should be carefully assessed during system design. This paper discuss
es a global approach for assessment of the dependability of industrial info
rmation systems, covering all the phases of a study, from initial system mo
deling to dependability analysis and evaluation. After the presentation of
the modeling concepts and tools, the paper focuses on the dependability eva
luation. Two complementary algorithms are discussed, one based on the devic
e stage method and the other on the automatic derivation of analytical expr
essions for the state probabilities and their symbolic evaluation. The cent
ral feature of these algorithms is their ability to deal with nonexponentia
l stochastic processes, a fundamental requirement for this class of systems
.