Performance of columnar database

Main Article Content

Mario Raul Morales-Morales https://orcid.org/0000-0002-7493-8072
Jhonatan W. Durán-Cazar http://orcid.org/0000-0002-8574-1435
Eduardo J. Tandazo-Gaona http://orcid.org/0000-0002-2209-3952
Santiago Morales Cardoso http://orcid.org/0000-0002-3833-9654

Abstract

Companies’ capacity to efficiently process a great amount of data from a great variety of sources anywhere and anytime is essential for them to succeed. Data analysis becomes a key strategy for most large organizations to get a competitive advantage. Hence, new issues should be considered when massive amounts of date are to be stored, because traditional relational database are not capable to lodge them. Such questions include aspects that range from the capacity to distribute and escalate the physical storage, to the possibility of using schemes or non-usual types of data. The main objective of this research is to evaluate the performance of the columnar databases in data analysis, comparing them with relational databases, to determine their efficiency using measurements in different test scenarios. The present study seeks to provide (scientific evidence) professionals interested in data analysis with a basic instrument for their knowledge, to include comparative tables with quantitative data that can support the conclusions of this research. A methodology of applied type and quantitative-comparative descriptive design is used, as it is the one of the most appropriate to study database efficiency characteristics. In the measurement, the method of averages is used for a number n of records, and it is supported in the Aqua Data Studio tool that guarantees a high reliability, as a specialized software for the administration of databases. Finally, it has been determined that the columnar databases have a better performance in data analysis environments.