On the Definition of Exception Handling Policies for Asynchronous Events in Workflow Processes

摘要

Exception handling during the execution of workflow processes is a frequently addressed topic in the literature. One important class of exceptions is those that represent predictable deviations from the normal behavior of the process that can be anticipated at design time. Such 'expected' exceptions are often caused by the occurrence of external events that are asynchronous with respect to the process. The desired exception handling response to these events will often depend on the current state of process execution. One important aspect of this stale is the relevant process instance's progress through the process model, which can be expressed in terms of the set of currently executing tasks. In this paper, we present a qualitative discussion on issues relevant to the definition of policies for handling asynchronous, expected exceptions. First, we highlight the requirement for workflow control data to be referenced in the policies if these exceptions are to be handled in a meaningful way. We then demonstrate that the definition of exception handling policies is not a trivial exercise in the context of complex processes, and discuss correctness criteria for these definitions. Finally, we outline a methodology for policy definition to ensure that the policy for each event is complete and consistent with respect to all possible states of process execution for the relevant process model.