A cache-based method to improve query performance of linked Open Data cloud

Abstract

The proliferation of semantic big data has resulted in a large amount of content published over the Linked Open Data (LOD) cloud. Semantic Web applications consume these data by issuing SPARQL queries. One of the main challenges faced by querying the LOD web cloud on account of the inherent distributed nature of LOD is its high search latency and lack of tools to connect the SPARQL endpoints. In this paper, we propose an Adaptive Cache Replacement strategy (ACR) that aims to accelerate the overall query processing of the LOD cloud. ACR alleviates the burden on SPARQL endpoints by identifying subsequent queries learned from clients historical query patterns and caching the result of these queries. For cache replacement, we propose an exponential smoothing forecasting method to replace the less valuable cache content. In the experimental study, we evaluate the performance of the proposed approach in terms of hit rates, query time and overhead. The proposed approach is found to outperform existing state-of-the-art approaches, increase hit rates by 5.46%, and reduce the query times by 6.34%.