A Comparative Analysis of NoSQL and SQL Databases: Performance, Consistency, and Suitability for Modern Applications with a Focus on IoT

Abstract

The rapid expansion of data-intensive technologies, such as big data analytics, cloud computing, real-time processing, and the Internet of Things (IoT), has pushed traditional relational database management systems (RDBMS) like MySQL beyond their limits, leading to the adoption of NoSQL databases such as MongoDB, Cassandra, Redis, and CouchDB. These NoSQL systems offer superior scalability, flexibility, and performance for managing large, diverse datasets, especially those generated by IoT environments, though they often sacrifice strict consistency and introduce complexities absent in SQL databases. This paper compiles findings from a broad range of recent studies to compare NoSQL and SQL databases across three critical aspects: performance, consistency models, and suitability for modern applications, with particular attention to IoT. Drawing on benchmarking tools like the Yahoo! Cloud Serving Benchmark (YCSB) and empirical research, the analysis examines their performance under various workloads, their approaches to data consistency, and their applicability to IoT, big data, and distributed systems. Results show that NoSQL databases excel in scalability and adaptability—key for IoT’s high-volume, unstructured data—while SQL databases maintain strengths in structured data management and transactional reliability. With insights into IoT-specific needs, such as MongoDB’s real-time processing and InfluxDB’s time-series capabilities, this study provides a practical framework for developers, researchers, and system architects to choose databases suited to specific application requirements, especially in the growing IoT field