MaskD : a tool for measuring masking fault-tolerance

Abstract

We present MaskD, an automated tool designed to measure the level of fault-tolerance provided by software components. The tool focuses on measuring masking fault-tolerance, that is, the kind of fault-tolerance that allows systems to mask faults in such a way that they cannot be observed by the users. The tool takes as input a nominal model (which serves as a specification) and its fault-tolerant implementation, described by means of a guarded-command language, and automatically computes the masking distance between them. This value can be understood as the level of fault-tolerance provided by the implementation. The tool is based on a sound and complete framework we have introduced in previous work. We present the ideas behind the tool by means of a simple example and report experiments realized on more complex case studies.