Comparative Performance Analysis of Four RDBMS Systems Integrated with Django's ORM

Abstract Object--Relational Mapping (ORM) frameworks make it easier to develop applications but often incur a significant performance overhead. This paper presents a comprehensive experimental evaluation of four popular relational databases (MySQL, PostgreSQL, Oracle, and Microsoft SQL Server) accessed via Django's ORM. Using the TPC-H benchmark, we measured query execution time, throughput, and resource utilization under varying query complexities and concurrency levels. PostgreSQL delivered the best overall performance under complex queries via the ORM, whereas Oracle’s superior raw SQL speed was largely offset by high ORM overhead. SQL Server showed balanced results, and MySQL performed well on simple queries but struggled with complex ones. We quantify the ORM-induced slowdown---ranging from 2.6 times in PostgreSQL to 5 times in Oracle---and demonstrate that proper indexing can improve performance by over 60% across all systems. These statistically significant findings underscore important trade-offs when using ORM frameworks. Our study provides new insights and practical guidance for selecting and tuning ORM-database combinations to achieve optimal performance.