Ensuring Consistency in Graph Cache for Graph-Pattern Queries

Wang, J., Ntarmos, N. and Triantafillou, P. (2017) Ensuring Consistency in Graph Cache for Graph-Pattern Queries. In: The 6th International Workshop on Querying Graph Structured Data (GraphQ '17), Venice, Italy, 21-24 Mar 2017,

133670.pdf - Published Version
Available under License Creative Commons Attribution Non-commercial No Derivatives.



Graph queries are costly, as they entail the NP-Complete subgraph isomorphism problem. Graph caching had been recently suggested, showing the potential to significantly accelerate subgraph/supergraph queries. Subsequently, Graph- Cache, the first full-fledged graph caching system was put forth. However, when the underlying dataset changes con- currently with the query workload proceeding, how to ensure the graph cache consistency becomes an issue. The current work provides a systematic solution to address this problem, by presenting an upgraded GraphCache system coined GraphCache+ (abbreviated as GC+). We develop two GC+ exclusive models that employ different approaches to deal with the consistency issue. Moreover, we present the logic of GC+ in expediting queries, bundled with the formally proved correctness. We evaluate the performance of GC+ by a real-world graph dataset and a number of query workloads with different characteristics, highlighting the considerable speedup in term of quantified benefit and overhead.

Item Type:Conference Proceedings
Glasgow Author(s) Enlighten ID:Triantafillou, Professor Peter and Ntarmos, Dr Nikos and Wang, Jing
Authors: Wang, J., Ntarmos, N., and Triantafillou, P.
College/School:College of Science and Engineering > School of Computing Science
Copyright Holders:Copyright © 2017 The Authors
First Published:First published in The 6th International Workshop on Querying Graph Structured Data (GraphQ '17)
Publisher Policy:Reproduced under a Creative Commons License
Related URLs:

University Staff: Request a correction | Enlighten Editors: Update this record