Advanced techniques for maintaining reliability of complex computer systems
Alvin Lim · 2002
Complex computer systems often involve complex coordination among distributed components that are susceptible to failure and reconfiguration during normal execution. New advanced techniques are required for maintaining the reliability of complex computer systems; traditional techniques used in conventional systems are based on assumptions that are invalid in complex computer systems. These advanced techniques permit complex interactions and exploit the semantics of complex applications to improve the efficiency of concurrency and recovery from failure. During failure recovery, global consistency is maintained by automatically analyzing dependencies among operation executions that are defined in terms of properties of the behavioral model. By analyzing application behavior, this technique provides the benefit of efficient recovery to a consistent intermediate state and reduces loss of computation or external work.