On Storage Policies for Semantic Web Repositories That Support Versioning

Abstract

Abstract. This paper concerns versioning services over Semantic Web (SW) repositories. We propose a novel storage index (based on partial orders), called POI, that exploits the fact that RDF Knowledge Bases (KB) have not a unique serialization (as it happens with texts). POI can be used for storing several (versionrelated or not) SW KBs. We discuss the benefits and drawbacks of this approach in terms of storage space and efficiency both analytically and experimentally in comparison with the existing approaches (including the change-based approach). For the latter case we report experimental results over synthetic data sets. POI offers notable space saving as well as efficiency in various cross version operations. It is equipped with an efficient version insertion algorithm and could be also exploited in cases where the set of KBs does not fit in main memory.

Evaluating query and storage strategies for RDF archives

“…Finally, [41] builds a partial order index keeping a hierarchical track of changes. This proposal, though, is a limited variation of delta computation and it is only tested with datasets having some thousand triples.…”

Semantic to intelligent web era

“…Preservation of CK over time, through: Versioning (maintaining different versions of the CK captured at different time stamps, hence allowing to maintain a history of CK variation and evolution over time) [113], Identifying the minimum amount of CK necessary for preservation, in order to safeguard pertinent data -following the users' needs, or perform certain specific tasks -following system requirements (such as intelligent search, storage, indexing, etc. ), instead of needlessly preserving the whole CK, hence drastically reducing recourse consumption.…”

Reliable Granular References to Changing Linked Data

“…Reproducibility and persistence have been ongoing concerns in this regard, as dataset identification and access has often been brittle and unreliable. Datasets based on Linked Data, as most types of datasets, are typically quite dynamic and change over time [36,9], and capturing the data's provenance [24] is crucial for their proper interpretation and reuse. Moreover, as we will show, scientific data analyses typically use relatively small subsets of Linked Data resources, but we currently lack reliable methods to refer to such subsets.…”