This paper addresses the dependability modeling of hardware and software fault-tolerant systems taking into account explicitly the dependency between the various components. These dependencies may result from functional or structural interactions between the components or from interactions due to global system reconfiguration and maintenance strategies. Since the resulting dependencies are stochastic, the dependability of the system cannot be evaluated by combination of the dependability of its components. This paper presents a framework for modeling the dependencies between the various components of a fault-tolerant system. Modeling is based on Generalized Stochastic Petri Nets (GSPNs). The modeling approach is modular: the behavior of each component and each interaction is represented by its own GSPN, while the system model is obtained by composition of these GSPNs. The composition rules are defined and formalized through clear identification of the interfaces between the component and the dependency nets. In addition to modularity, the formalism brings flexibility and re-usability thereby allowing for easy sensitivity studies with respect to the assumptions that could be made about the behavior of the components and the resulting interactions. This approach is applied to a simple, but still representative, example.