Treffer: A REVIEW ON OBJECT-ORIENTED, OBJECT-RELATIONAL, RELATIONAL, AND NOSQL DATABASES: EVOLUTION, INTEGRATION, AND PERFORMANCE TRADE-OFFS

This systematic review evaluates the architectural principles, integration challenges, and performance trade-offs of Object-Oriented (OO), Object-Relational (OOR), Relational (SQL), and NoSQL databases, focusing on their evolution, interoperability, and benchmarking outcomes. The study adopts the PRISMA 2020 framework to ensure methodological rigour, with a comprehensive search across IEEE Xplore, ACM Digital Library, Scopus, ScienceDirect, and SpringerLink. Following the PRISMA process as a methodology used, 243 initial records were identified, 99 duplicates removed, 95 excluded during title and abstract screening, 29 excluded after full-text review, and 2 excluded for being published before 2020, resulting in 18 studies retained for final synthesis. Architectural analysis reveals that OO-DBMSs improve developer productivity by up to 25% in iterative builds but suffer from limited interoperability. OORDBMSs balance object and relational paradigms with a 10–15% storage overhead, SQL databases maintain <50 ms query latencies on billion-row datasets, and NoSQL achieves high ingestion rates with 15–20% data duplication. Performance modelling using regression and complexity analysis indicates statistically significant differences (p < 0.05) in execution time, query complexity, and memory usage across workloads, highlighting trade-offs between scalability, consistency, and feature richness. The review proposes a unified performance evaluation framework and practical integration guidelines for hybrid database adoption, addressing identified gaps such as the lack of standardised cross-model benchmarking. Future research should explore longitudinal adoption patterns and quantum-inspired architectures to address emerging scalability and optimisation demands.