An architecture and a methodology for recording provenance information and creating provenance-capable applications was developed in the Grid Provenance EU project. In this context, “provenance information” of data refers to information on the process that resulted in these data. This information is important or even indispensable for a number of applications, e.g. in order to have a reliable documentation of the individual steps performed in engineering calculations, or for secure information on the location of and changes to data in medical data processing and management. Therefore, the demonstration applications in the provenance project included a complex manoeuvre simulation from the aeronautics field, an organ transplant management application and a patient data management application from the field of medicine. 

The architecture developed in the provenance project was designed as a Service-Oriented Architecture. The provenance information to be recorded from the application is saved in a “provenance store” and can be retrieved and evaluated later. A programming library (“client side library”) was developed in Java for simple recording of this information from within an application. The following illustration shows the provenance architecture:

In the provenance project, a workflow management system for complex engineering simulations was extended at DLR to include provenance recording, and evaluated using various workflows (parameter variation and simulation of complex flight manoeuvres) [PDF]. The resulting knowledge is applied in the AeroGrid project.

In addition to the existing Java library, DLR is currently developing a client-side provenance library in Python, enabling its integration into a number of applications, particularly the DataFinder.