Policy-Driven System Configuration for Dependability

When managing a complex information system for critical activities, reaching an high level of dependability is a key requirement. A central challenge is identifying a set of alternative system configurations to be enforced when particular system situations take place. In large systems, this task requires automation as the administrator can hardly cope with the complexity of all components and their interactions. We propose an innovative approach, based on model transformation, to perform automatic generation of alternative configurations taking into account dependability constraints.