Comparative Analysis of Relational (Oracle) and Non-Relational (Cassandra) Databases for Business Intelligence

AKINMOLADUM, T.M, LAKE, Peter, SAMUEL, O.W. and DOMDOUZIS, Konstantinos (2017). Comparative Analysis of Relational (Oracle) and Non-Relational (Cassandra) Databases for Business Intelligence. International Journal of Multidisciplinary and Current Research (IJMCR), 5.

Domdouzis Comparative Analysis of Relational (Oracle) and Non-Relational (Cassandra) Databases.pdf - Published Version
Creative Commons Attribution.

Download (834kB) | Preview
Official URL:


The need for business intelligence systems (BI) cannot be overemphasised because of the huge data constantly being generated in the daily operations of business organisations and the opportunity provided to discover new insights for the improvement of organisational effectiveness and efficiency from the data. This study attempts to carry out performance related tests on Oracle and Cassandra in order to propose a suitable database for business intelligence. Firstly, the extract, transform and load (ETL) processes was used to move data into Oracle and Cassandra virtual machines. Secondly, SQL and NoSQL queries were run on the data in three iterations to test for performance in selected workloads (Create and load process, read, update, delete and join operations) both before and after query optimisation. To create a common ground for comparison, similar queries were run on similar datasets on both databases. Then the results from the tests were statistically analysed using Microsoft Excel. Experimental results show that the latency values of Oracle are observed to be lower than that of Cassandra, accuracy values of Cassandra are observed to be nearly the same with that of Oracle in the create and load process, while their accuracy values are observed to be slightly different in the remaining tested workload, and the throughput values of Cassandra are observed to be higher than that of Oracle. Also, the extent to which these performance outcomes support data analytics for BI is hereby presented.

Item Type: Article
Departments - Does NOT include content added after October 2018: Faculty of Science, Technology and Arts > Department of Computing
Depositing User: Konstantinos Domdouzis
Date Deposited: 16 May 2017 10:24
Last Modified: 18 Mar 2021 15:33

Actions (login required)

View Item View Item


Downloads per month over past year

View more statistics