Provenance Tracking in UNICORE

The automated tracking and storage of provenance information allows users of scientific workflow systems to validate and reproduce results of their experiments. Until now, UNICORE has not been providing comprehensive provenance features. For this reason, we plan to equip UNICORE with a flexible provenance tracking mechanism. Our goal is to ensure a suitable traceability of job and workflow processes into a description format supporting query capabilities and interoperability. In a first step, we have started a detailed analysis investigating existing provenance standard formats, tools, and ontologies. In addition, we defined the pieces of data and information that need to be tracked by the integrated provenance system. In the next step, we are examining solutions for the persistent storage of provenance data. In this context, we distinguish provenance tracking metadata from input and result data of computations. Whereas for the former a suitable database system is required, we consider the use of version control systems to keep track of the latter. All these preliminary considerations and investigations will lead to the implementation of the UNICORE Provenance Tracking System