Home // DBKDA 2020, The Twelfth International Conference on Advances in Databases, Knowledge, and Data Applications // View article

A Graph Database Storage Engine for Provenance Graphs

Authors:
Changhong Liu
Hancong Duan

Keywords: Graph Database; Graph Analytics and Storage; Provenance Graphs.

Abstract:
The rapid development of high-speed networks has created a massive amount of data. Storing and mining such data is of great research value. Knowledge graphs and graph databases have widely been studied and applied as an effective means to mine the associated data in the past few years. Provenance graphs provide powerful ways to observe the changes in a graph, especially in graph analysis. The update operation will produce massive provenance graphs from a given graph as time goes on. It is a challenge to store and query these massive provenance graphs efficiently. Meanwhile, the query performance itself must be guaranteed. To address this challenge, this paper presents a graph database storage engine called T-GDB (Temporal dimension - Graph Database). This system binds the topology of the graph to each vertex in the graph and rebuilds the graph in real-time when analyzing the graph. T-GDB can analyze the changes in a graph over time and can also access the provenance of the specified graph through the index tree. T-GDB can support these application scenarios such as the knowledge reasoning of knowledge graphs and the information mining for specified graphs. This paper describes the format of data storage, the index, and the implementation of this system. Finally, this paper compares the proposed graph database storage engine to several existing mainstream graph databases to verify the feasibility and efficiency of this design. Our experimental results demonstrate that the proposed graph database storage engine has better performance and more efficient graph analysis than existing methods.

Pages: 1 to 6

Copyright: Copyright (c) IARIA, 2020

Publication date: September 27, 2020

Published in: conference

ISSN: 2308-4332

ISBN: 978-1-61208-790-0

Location: Lisbon, Portugal

Dates: from September 27, 2020 to October 1, 2020