Ultra-Dependable Real-Time Computing

Abstract : The research effort this year focused in four major areas. (1) Developing schedulable fault recovery strategies that guarantee minimum interference with fault free tasks, and graceful degradation; (2) Developing analytical techniques that support quantitative analysis of computer architectures for fault-tolerant computation; (3) Investigating how intelligent integrated agents can enhance distributed system' dependability; and (4) Developing network diagnostic strategies with provable timing properties. The project members are also concerned with the transition of the theory being developed to the user community, especially the U.S. Navy. For this reason, project members have a close relationship with the RTSIA (Real-Time Scheduling in Ada) project in the CMU Software Engineering Institute, and Distributed Combat Control Project at Naval Ocean Systems Center in San Diego.