Dealing with Logical Failures for Collaborating Workflows

Logical failures occurring during workflow execution require the dynamic adaptation of affected workflows. The consequences such a dynamic adaptation may have for collaborating workflows have not yet been investigated sufficiently. We propose a rule-based approach for dynamic workflow adaptation to deal with logical failures. In our approach, workflow collaboration is based on agreements specifying the delivery time and quality of objects a workflow expects from its collaboration partners. Our mechanisms decide which collaborating workflows have to be informed when a dynamic adaptation is performed. In particular, we estimate the temporal and qualitative implications a dynamic adaptation has for collaboration partners. Because of the automated handling of logical failures, we expect that our approach significantly improves the robustness and correctness of collaborating workflows. The approach has been developed in the context of collaborative workflow-based care for cancer patients.