Managing Provenance in Scientific Workflows with ProvManager

Running scientific workflows in distributed environments is motivating the definition of provenance gathering approaches that are loosely coupled to the workflow systems. We have proposed a provenance gathering strategy that is independent from workflow system technology. This strategy has evolved into a provenance management system named ProvManager. The main principle is that each workflow activity should collect its own provenance data and publish them in a repository which scientists can access to make their queries. In this paper we show how provenance is captured along distributed heterogeneous systems. Two main strategies are used to capture provenance: using Prolog predicates to register provenance, and using an API for the communication between the wrapped activity and the ProvManager.