Process Dependencies and Process Interference Rules for Analyzing the Impact of Failure in a Service Composition Environment

This paper presents a process dependency model for dynamically analyzing data dependencies among concurrently executing processes in an autonomous, distributed service composition environment. Data dependencies are derived from incremental data changes captured at each service execution site. Deltas are then used within a rule-based recovery model to specify how failure recovery of one process can potentially affect another process execution based on application semantics. This research supports relaxed isolation and application-dependent semantic correctness for concurrent process execution, with a unique approach to resolving the impact of process failure recovery on other processes, using data dependencies derived from distributed, autonomous services.