Comparative Analysis of Relational and Non-Relational Database Models: A Case Study on Travel Booking Systems

Aim/Background

The study aims to evaluate and compare relational and non-relational (NoSQL) database models in the context of travel booking systems. Traditional relational databases like MySQL are known for their data integrity and structured schema, while NoSQL models like ArangoDB, Apache Cassandra, Memcached, and MongoDB offer scalability and flexibility. This research investigates which model performs best across various database operations relevant to real-world scenarios.

Methodology

The research used both theoretical and practical methods. A literature review was conducted using sources such as Google Scholar, IEEE Xplore, and database documentation. The team implemented a travel booking system to perform standardized operations like insert, update, delete, retrieve, access control, and integrity checks using five selected DBMSs. Performance metrics included execution time, CPU usage, throughput, and constraint handling, and were analyzed using visualizations and comparative tables.

Results

Apache Cassandra showed the best overall performance for flight booking systems due to high throughput and scalability. MySQL performed exceptionally in data integrity and consistent performance under constraints. ArangoDB showed flexibility and low insertion time but required a learning curve. Memcached excelled in insert/delete throughput but lacked persistence. MongoDB offered schema flexibility but lagged in constrained scenarios and multi-document transactions.

Discussion

The findings reveal that database selection should align with specific application needs. NoSQL models are preferable for high-speed, flexible operations, while relational databases are superior for structured, integrity-focused use cases. The study corroborates prior research advocating NoSQL for big data applications, yet highlights the enduring value of RDBMSs in mission-critical systems.

Conclusion

Apache Cassandra is the most appropriate choice for large-scale, dynamic systems like travel booking due to its high availability and performance. MySQL and ArangoDB also have strong points for specific tasks. The research emphasizes that no one-size-fits-all model exists, and DBMS selection must consider operation type, data size, and performance requirements.