Provenance Tracking with Bit Vectors

Abstract

Data provenance (information about the origin of data items and the transformations that they have gone through) can be a useful security tool, particularly for forensic investigation. Provenance information can also be used to enforce information assurance concepts like integrity and authenticity. The current techniques for finding data source and lineage information are annotations and query inversion. Annotations lack scalability and require additional space for storing and querying provenance information, while query inversion incurs more time and processing overhead. In this paper, we offer fast and scalable models for computing source information, each of which are based on associating bit vectors with data sources.