A model driven approach for assessing survivability requirements of critical infrastructures

Critical infrastructures are complex networked systems. They must be able to provide essential services, even when they are compromised by intentional or accidental threats. Guaranteeing essential services means to ensure survivability with an adequate Quality of Service (QoS). This paper proposes a model-driven approach for the assessment of survivability requirements. In particular, we propose a graphical Survivability Assessment Model (SAM), based on UML. It is automatically derived from a UML specification that encompasses essential services, service modes, threats and survivability strategies. Furthermore, model-driven techniques are used to assess the SAM. Then, we propose some preliminary property verifications to discover flaws in the specification. The model driven paradigm is used to ensure high level of usability and abstraction of the artifacts that are key issues in communication among stakeholders. The approach has been applied to a scaled-down model of a smart grid, an evolution of traditional power grids based on high performance and dependable computer networks.