A Framework for Model-driven Scientific Workflow Engineering

Scientific workflows are one important means in the context of data-intensive science for reliable and efficient scientific data processing in distributed computing infrastructures such as Grids. A common trend is to adapt existing and established business workflow technologies instead of developing own technologies from scratch. This thesis provides a model-driven approach for scientific workflow engineering, in which domain-specific languages (DSLs) tailored for a certain scientific domain are used for scientific workflow modeling, and automated mapping techniques for technical execution are developed and evaluated. The Business Process Model and Notation (BPMN) is thereby used at the domain-specific layer and the Web Services Business Process Execution Language (BPEL) at the technical layer. The implementation uses the Eclipse Modeling Framework (EMf) and is evaluated in three application scenarios.