Efficient Provenance Storage for RDF Dataset in Semantic Web Environment

Over the web, excessive growth of RDF (Resource Description Framework) data leads the data authentication issues such as trouble in trusting data and verifying data qualities. Oftentimes, data into the web are published along with provenance for measuring trust and quality. Provenance provides authentication information, like, who modified and published the data, from where the data are originated and what manipulations have been applied to the data. During the last decade, various approaches have been evolved for representing and storing provenance information of RDF datasets. Most of them have followed the annotation-based approach. The annotation-based approach suggests that the metadata should be salted away in separate documents to maintain a rich set of provenance. While doing this, it demands more space for storing coarse-grained and fine-grained provenance. Nevertheless, they do not consider the techniques for redundancy reduction for duplicate provenance values exist in the dataset, in particular, when provenance is defined at the instance level. As a consequence, the size of the provenance may outsize the actual data size itself. In such instances, there should be provisions for reducing the space occupied by the duplicate provenance values. In this article, an approach has been proposed to store provenance information efficiently based on inheritance techniques. The provenance is pre-computed during the generation of the RDF data. It is then stored in an optimized way while minimizing the storage area for repeated provenance values. The result is quite promising in the sense that, the storage size is reduced considerably, as compared to the normally computed provenance.