This presentation is about the preliminary results for the development of a
Linked Data Fragment Server - built using Python and Redis - for providing a
lightweight mechanism for retrieving cached RDF triples for use by Semantic Web
applications.
The Problem
Providing a full SPARQL endpoint for both small and large RDF linked-data sets is
costly both in time and resources, especially for more complex queries. This
often results in poor responses and unreliable service, especially if the SPARQL
endpoint trying to handle multiple clients querying the RDF graph database.
A Promising Alternative
Ruben Verborgh of Ghent University in
Belgium originated the concept of Linked Data Fragments
that offers a middle-ground between
different options for accessing RDF graph data. Instead of needing high-powered server to handle
the processing load required for hosting a full SPARQL endpoint or the alternative of just
providing a data dump where all of the data is processed locally to be useful, the Linked
Data Fragments approach
instead offers a lightweight querying pattern called a
Triple Pattern Fragment that returns one or more triples based on a simplified
syntax.