Performance Impact of Parallel Access of Time Series in the Context of Relational, NoSQL and NewSQL Database Management Systems

Sebastian Pritz, Martina Zeinzinger, Christoph Praschl, Oliver Krauss, Martin Harrer

Research output: Chapter in Book/Report/Conference proceedingsConference contributionpeer-review

Abstract

Time series data is generated in various application areas, such as IoT devices or sensors in vehicles. This type of data is often characterized by a high resource demand due to the interval at which information is measured ranges from daily down to milliseconds. Next to the frequency, the number of data sources, for example hundreds of sensors in modern airplanes generating time series concurrently, is typical for such big data scenarios. Such scenarios require the persistence of the measurements for further evaluations. In this work, we introduce an artificial data benchmark for relational, NoSQL, and NewSQL database management systems in the context of time series. We compare these databases by having multiple read and write data sources accessing the database management systems simultaneously. The evaluation shows that no tested system outperforms all other systems. While DolphinDB shows the highest read performance in single-user scenarios, CrateDB is able to show its advantages regarding when multiple users access the data simultaneously.

Original languageEnglish
Title of host publicationInternational Conference on Electrical, Computer, Communications and Mechatronics Engineering, ICECCME 2023
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9798350322972
DOIs
Publication statusPublished - Jul 2023
Event2023 International Conference on Electrical, Computer, Communications and Mechatronics Engineering, ICECCME 2023 - Tenerife, Canary Islands, Spain
Duration: 19 Jul 202321 Jul 2023

Publication series

NameInternational Conference on Electrical, Computer, Communications and Mechatronics Engineering, ICECCME 2023

Conference

Conference2023 International Conference on Electrical, Computer, Communications and Mechatronics Engineering, ICECCME 2023
Country/TerritorySpain
CityTenerife, Canary Islands
Period19.07.202321.07.2023

Keywords

  • Big Data
  • Database
  • IoT
  • Sensor Networks
  • SQL
  • Time Series

Fingerprint

Dive into the research topics of 'Performance Impact of Parallel Access of Time Series in the Context of Relational, NoSQL and NewSQL Database Management Systems'. Together they form a unique fingerprint.

Cite this