Error recovery mechanisms for remote procedure call-based systems

The authors present a set of mechanisms for implementing application-dependent and target-dependent error recovery policies without changing the application designed to face only a particular (usually empty) set of failures. These mechanisms are based on the assumption that the program is constituted by a set of resources communicating through procedural interfaces. The resulting environment allows the implementation and the integration of well-known approaches as exception handling, recovery blocks, dynamic reconfiguration, and atomic actions.<
>