A Policy-based Dependability Management Framework for Critical Services

Abstract—Many critical activities rely on the correct and uninterrupted operation of networked Computer Information Systems (CIS). Such systems are however exposed to many different kinds of risk, and thus many researches have been taking place for enabling them to perform self-monitoring and self-healing, and so maintaining their operation over time as specified by domain policies. These capabilities are the basis of what is commonly referred to as dependability. The DESEREC project has defined a tiered architecture as a policy based framework to increase the dependability of existing and new networked CIS, using technology-independent information which is translated at runtime to suit the managed components. This paper delves into how DESEREC builds and manages large critical systems through an agent-based distributed framework, and how it is able to respond to any adversity effectively, such as intrinsic failures, misbehavior or malicious internal use, and attacks from the outside. An illus-trative example is used throughout this paper to demonstrate all the concepts and definitions presented