The Suitability of Graph Databases for Big Data Analysis: A Benchmark

This publication doesn't include Faculty of Arts. It includes Institute of Computer Science. Official publication website can be found on muni.cz.

Authors

MACÁK Martin ŠTOVČIK Matúš BÜHNOVÁ Barbora

Year of publication 2020
Type Article in Proceedings
Conference Proceedings of the 5th International Conference on Internet of Things, Big Data and Security - Volume 1: IoTBDS
MU Faculty or unit

Institute of Computer Science

Citation
Web https://www.scitepress.org/PublicationsDetail.aspx?ID=qc6Zz7Qsgn0=&t=1
Doi http://dx.doi.org/10.5220/0009350902130220
Keywords Big Data; Benchmark; Graph Database; Neo4j; PostgreSQL
Description Digitalization of our society brings various new digital ecosystems (e.g., Smart Cities, Smart Buildings, Smart Mobility), which rely on the collection, storage, and processing of Big Data. One of the recently popular advancements in Big Data storage and processing are the graph databases. A graph database is specialized to handle highly connected data, which can be, for instance, found in the cross-domain setting where various levels of data interconnection take place. Existing works suggest that for data with many relationships, the graph databases perform better than non-graph databases. However, it is not clear where are the borders for specific query types, for which it is still efficient to use a graph database. In this paper, we design and perform tests that examine these borders. We perform the tests in a cluster of three machines so that we explore the database behavior in Big Data scenarios concerning the query. We specifically work with Neo4j as a representative of graph databases and PostgreSQL as a representative of non-graph databases.
Related projects: