Statistical Description of PostgreSQL and Cassandra in CRUD Operations

In recent years, the increase in the number of users and applications on the Internet has generated a problem of storing large amounts of information and continuing to manage it more efficiently with high performance and availability rates. In this context, databases are the most efficient tools for storing information regardless of business need, quantity, or data type. Relational databases such as PostgreSQL have demonstrated maturity, reliability, and data integrity; on the other hand, Cassandra has been a fault-tolerant database for the past few years when handling large amounts of data. However, there is a constant question about whether relational databases are better at meeting these challenges than non-relational databases. This paper aims to evaluate the performance of PostgreSQL and Cassandra database managers in CRUD operations. Therefore, a comparative study is conducted to analyze the performance in writing, reading, and data processing when handling CRUD operations, following the YCSB framework that supports linear (PostgreSQL) and non-linear (Apache Cassandra) database managers. The results show that the linear regression models explain 98.73%, 72.56%, and 77.54% of the variability between the databases.