The Failure Recovery Management Framelet

Abstract

There is usually no standardized approach to failure recovery in present AOCS applications. Responses to a failure are coded in dedicated procedures that are called at the point where the failure is detected. This framelet proposes a more general solution by separating the management of the failure recovery process from the implementation of specific failure recovery counter-measures. The latter are obviously application-specific and therefore are treated as a framework hot-spot. The failure recovery function is implemented by a failure recovery manager based on the manager meta-pattern. The failure detection manager is one of the framework functionality managers. Its job is to react to failure reports (as they would be typically generated by the failure detection function) by triggering the execution of failure recovery strategies.