Efficient Mobile Querying of Distributed RDF Sources

Recent advancements in mobile devices and the omnipresence of wireless connectivity have turned handheld devices into powerful mobile web clients. This evolution has made the wealth of online Semantic Web data accessible to mobile devices. To support this new opportunity, we present a semantic technology-based client-side solution to efficiently query large amounts of distributed online RDF sources. The main idea is to continuously extract and manage metadata from RDF sources and to store this metadata in a local Source Index Model (SIM). At query execution time, the SIM is consulted to identify and assemble potentially relevant sources. In contrast to other approaches [1-3], we do not rely on any query endpoints. Our solution pays attention to the computational limitations of mobile devices by querying only relevant sources and thereby improving the overall query execution time. We do not aim to replace existing mobile query engines, but rather build on them to efficiently and transparently query large sets of online RDF sources. The first step of our solution involves the creation of a SIM as highlighted in Fig. 1. Each time we receive a new source reference from an application, the source is downloaded and relevant metadata is extracted and stored in the SIM. Since RDF is a predicate-based formalism, we base our approach on the presence of predicates in the data sources to filter out irrelevant data sources similar to [1,2]. We further aim to exploit the semantic information embedded in RDF documents in terms of the type of the predicate subjects (i.e. domain). RDF sources and SPARQL queries often use this domain information to detail triples and restrict triple patterns. We consider two different SIM variants: the first variant stores only the found predicates while the second variant manages the found predicates together with their domains. Note that the second SIM variant still supports queries with unspecified predicate domains. The second step of our approach consists in resolving an application query over the combined data set of relevant sources as shown in Figure 2. First, the query analyzer extracts the metadata of the given SPARQL query that is compatible with the data stored in the SIM (predicates or predicates with their domains). The triple patterns occurring in the WHERE clause of a SPARQL query are used to restrict the query results in the graph pattern matching process. Therefore, we need to examine the WHERE clause in order to identify …