A theory for observational fault tolerance

In general, faults cannot be prevented; instead, they need to be tolerated to guarantee certain degrees of software dependability. We develop a theory for fault tolerance for a distributed pi-calculus, whereby locations act as units of failure and redundancy is distributed across independently failing locations. We give formal definitions for fault tolerant programs in our calculus, based on the well studied notion of contextual equivalence. We then develop bisimulation proof techniques to verify fault tolerance properties of distributed programs and show they are sound with respect to our definitions for fault tolerance.