EXPERIMENTAL PERFORMANCE EVALUATION OF MONGODB AND SQL SERVER UNDER LARGE-SCALE WORKLOADS

Abstract

The rapid growth of large-scale and heterogeneous data generated by web applications, cloud platforms, and Internet of Things (IoT) systems has increased the need for efficient and scalable data management solutions. Traditional relational database management systems (RDBMS), such as Microsoft SQL Server, ensure strong consistency and data integrity, while NoSQL systems, like MongoDB, provide schema flexibility and horizontal scalability. Selecting an appropriate database architecture remains a critical design decision for modern applications. This study presents a controlled experimental performance evaluation of Microsoft SQL Server and MongoDB under identical deployment conditions. Both systems were containerized using Docker and tested with standardized datasets ranging from 10K to 5M records. Performance was assessed in terms of insertion time, query latency, update and delete execution time, CPU utilization, memory consumption, and scalability behavior. Monitoring was conducted using Prometheus and Grafana to capture system metrics. Experimental results indicate that MongoDB shows better performance in insert operations, queries, and resource efficiency, while SQL Server shows advantages in structured and type-based queries. The findings highlight that database selection should be driven by workload characteristics and application requirements rather than general performance assumptions.