Model-Driven Engineering of Dependable Systems

Abstract

Improving the dependability of a computer system increases the acquisition cost so much that many systems are built without a cost-effective level of dependability. This motivates our decision to work on reducing the development effort and competence required to create dependable, distributed, reactive systems. The scope is narrowed to extending the SPACE method with software-implemented fault-tolerance mechanisms and providing tool-supported fault removal in the form of model checking. The results so far mainly cover fault removal, but we also have some early results on providing fault-tolerance mechanisms at the application layer. We discuss future work as well.