On large and complex distributed systems hardware and software faults, as well as vulnerabilities, exhibit significant dependencies and interrelationships. Being able to assess their actual impact on the overall system dependability is especially important. The goal of this paper is to propose a unifying way of describing a complex hardware and software system, in order to assess the impact of both vulnerabilities and faults by means of the same underlying reasoning mechanism, built on a standard Prolog inference engine. Some preliminary experimental results show that a prototype tool based on these techniques is both feasible and able to achieve encouraging performance levels on several synthetic test cases.