Due to the decentralised and autonomous architecture of the Web of Data, data replication and local deployment of SPARQL endpoints is inevitable. Nowadays, it is common to have multiple copies of the same dataset accessible by various SPARQL endpoints, thus leading to the problem of selecting optimal data source for a user query based on data properties and requirements of the user or the application. Quality of Service (QoS) parameters can play a pivotal role for the selection of optimal data sources according to the user's requirements. QoS parameters have been widely studied in the context of web service selection. However, to the best of our knowledge, the potential of associating QoS parameters to SPARQL endpoints for optimal data source selection has not been investigated. In this paper, we define various QoS parameters associated with the SPARQL endpoints and represent a semantic model for QoS parameters and their evaluation. We present a monitoring service for the SPARQL endpoint which automatically evaluates the QoS metrics of any given SPARQL endpoint. We demonstrate the utility of our monitoring service by implementing an extension of the SPARQL query language, which caters for user requirements based on QoS parameters and selects the optimal data source for a particular user query over federated sources.
展开▼
