Modeling and Querying Scientific Workflow Provenance in the D-OPM

We present the D-OPM, a model that extends the Open Provenance Model (OPM) with workflow-specific aspects. In particular, our model captures aspects such as the workflow structure, traces, data structure, and workflow evolution. Thus, it enables scientists to obtain detailed information about the origin of data resulting from past experiments, as well as about the process itself and its possible future executions. A reference implementation of the D-OPM validates our model and opens the opportunity for interoperation with multiple workflow systems. Furthermore, to facilitate querying D-OPM data we introduce a querying mechanism based on regular path queries (RPQs) on provenance graphs. Our RPQs evaluator is built on a relational DBMS which makes it robust and extensible.