Assembling total provenance from distributed systems

Nicholas Car

The PROV data model, and several of its predecessors (ODM, PML) are graph-based meaning they are suitable for schema-less use. In addition, the use of Linked Data technologies by PROV and these predecessors mean that it is able to be used for provenance from distributed sources: provenance of an object of interest is able to be, and perhaps expected to be, stored in multiple places rather than in a single place.

While the theoretical use of distributed provenance is well described, few practical guides exist to assist implementations. This Pattern describes several methods to assemble provenance from distributed systems into a single report.