Rapid Deterministic Fault Detection in Distributed Systems

The primary objective of this project is to develop rapid,deterministic and standards-compliant fault detection strategies andmechanisms for complex distributed systems, in particular, for theNavy's Open Architecture. The project will develop a fault monitoringplatform that performs health checks with several degrees of rigor,and that adapts the fault monitoring frequencies and timeouts to thecurrent load on the system. It will exploit the applications' requestand reply messages for fault monitoring within the context of areplication and recovery infrastructure that provides real-time faulttolerance. The project will also investigate a hierarchical faultdetection, analysis and notification infrastructure that providesaccurate and timely characterizations of faults and damage in complexdistributed systems. The anticipated benefits and potential commercial applications include:* Rapid, and accurate, detection and analysis of faults in complex distributed systems.* Fault detection mechanisms that are easy to use and easy to integrate into commercial system management and resource management products.* Fault detection mechanisms for commercial distributed systems, including telecommunications, industrial control and manufacturing, transportation and medical systems.* Rapid, accurate and standards-compliant fault detection mechanisms for the Navy's Open Architecture, as well as for other defense and Government systems.