A Cache-based Optimizer for Querying Enhanced Knowledge Bases

TL;DR

This paper introduces a cache-based optimizer for SPARQL query systems that significantly enhances query performance on large knowledge bases like DBpedia and Linked GeoData by prefetching and caching predicted results.

Contribution

The paper presents a novel cache-based optimization scheme for SPARQL query systems that outperforms existing systems in query speed through prefetching and caching strategies.

Findings

Query performance improved over state-of-the-art systems

Effective caching reduces query latency

System tested on DBpedia and Linked GeoData

Abstract

With recent emerging technologies such as the Internet of Things (IoT), information collection on our physical world and environment can be achieved at a much higher granularity and such detailed knowledge will play a critical role in improving the productivity, operational effectiveness, decision making, and in identifying new business models for economic growth. Efficient discovery and querying such knowledge remains a key challenge due to the limited capability and high latency of connections to the interfaces of knowledge bases, e.g., the SPARQL endpoints. In this article, we present a querying system on SPARQL endpoints for knowledge bases that performs queries faster than the state-of-the-art systems. Our system features a cache-based optimization scheme to improve querying performance by prefetching and caching the results of predicted potential queries. The evaluations on query sets from SPARQL endpoints of DBpedia and Linked GeoData showcase the effectiveness of our approach.