Knowledge Graph Management on the Edge

TL;DR

This paper introduces SuccinctEdge, an in-memory RDF store optimized for edge computing, enabling efficient SPARQL querying and reasoning without decompression, suitable for low-latency data management.

Contribution

It presents a novel, compact, decompression-free RDF store tailored for edge environments, enhancing data integration and reasoning capabilities.

Findings

Demonstrates high efficiency on real-world datasets

Supports reasoning with ontology in-memory

Achieves low-latency query responses

Abstract

Edge computing emerges as an innovative platform for services requiring low latency decision making. Its success partly depends on the existence of efficient data management systems. We consider that knowledge graph management systems have a key role to play in this context due to their data integration and reasoning features. In this paper, we present SuccinctEdge, a compact, decompression-free, self-index, in-memory RDF store that can answer SPARQL queries, including those requiring reasoning services associated to some ontology. We provide details on its design and implementation before demonstrating its efficiency on real-world and synthetic datasets.