An Experimental Study on the Impact of Database Normalization on Query Performance and Storage Efficiency in Relational Database Systems

Normalization of the database is a basic database management tool that seeks to reduce data redundancy, enhance data integrity, and ensure data structures are consistent. Despite the significance of normalization in the logical organization of the database, the impact of normalization on the performance of the database is a critical issue. This research seeks to examine the effect of database normalization on the performance of a relational database. An experimental model has been proposed to examine the effect of database normalization on the performance of a database. The experimental model has been used to examine database structures created with different levels of normalization, including Unnormalized Form (UNF), First Normal Form (1NF), Second Normal Form (2NF), and Third Normal Form (3NF). A transaction database has been created to mimic a real-world scenario of an online retail management application. Several performance metrics such as query execution time, storage usage, join operation complexity, and CPU resource usage are measured and evaluated for various schema structures. From the experimental results, it is clear that as the level of normalization increases, data redundancy and storage usage are minimized, and data consistency is also maximized. Nevertheless, as the level of normalization increases, the number of join operations also increases. The increase in join operation might affect the query execution time to some extent. The findings of this research are useful to database designers and administrators to select an appropriate level of normalization for real-world database applications.