Mapping Semantic Workflows to Alternative Workflow Execution Engines

Current workflows have limited reusability and shelf life because they specify particular application software to be run at each step, which can become obsolete or no longer run. We describe semantic workflow representations composed of domain tasks, which the system then automatically maps to the application codes available in the particular execution environment. We have implemented this approach in the WINGS semantic workflow system, and demonstrate the mapping of workflows to different sets of application codes depending on the workflow execution engine and execution environment selected.