An Intermediate Dependability Model for state-based dependability analysis

Model-Driven engineering (MDE) aims to elevate models in the engineering process to a central role in the speciication, design, integration, validation, and operation of a system. MDE is becoming a widely used approach within the dependability domain: the system, together with its main dependability-related characteristics, is represented by engineering language models, while automatic transformations are used to generate the analysis models for the dependability analyses. An intermediate model eases the definition of transformations, providing an additional abstraction layer, through a representation that is independent both of the high-level modeling language and of the specific analysis formalism. In this work we define a new Intermediate Dependability Model (IDM) to support state-based dependability analysis. With respect to previous approaches based on intermediate models we introduce new modeling features; more in detail, we provide support for the modeling of detailed fault/error/failure chains inside component, complex propagation paths, internal error propagation and error compensation, multiple failure modes, error detection activities, maintenance activities, and a more detailed specification of the measures of interest that should be evaluated. In order to provide a better understanding of how IDM modeling elements are used we also provide some simple modeling examples using a graphical representation. Finally, model transformation rules are provided in order to automatically derive an IDM model starting from models built using a prototype of the CHESS ML language, an engineering modeling language based on UML developed in the ongoing CHESS project.