The aim of this thesis is to analyze and experimentally compare selected Java frameworks for working with relational databases (Hibernate, EclipseLink, DataNucleus, and MyBatis) with respect to data access efficiency and scalability for larger volumes of data (ranging from thousands to millions of records).

The thesis focuses on comparing ORM frameworks that implement the Jakarta Persistence specification (Hibernate, EclipseLink, DataNucleus) with the SQL mapper MyBatis. The evaluation will primarily consider performance in terms of throughput and response time when executing database operations on entities of varying complexity and with different types of relationships between entities (e.g., foreign keys, inheritance).

As part of the thesis, the selected frameworks will be briefly introduced, a methodology for measuring their performance and scalability will be designed, and experiments will be conducted on datasets of different sizes. Measurements will be performed both on a defined transaction mix inspired by OLTP workloads and on the level of individual transactions. Based on the obtained results, the strengths and weaknesses of the individual frameworks will be evaluated, and recommendations for their use will be formulated with respect to different workload characteristics and application requirements.

The results, including any associated code, will be publicly available in the IS under the MIT License.