Hybrid SQL/NoSQL Architecture for Optimizing the Performance of IoT Air Quality Monitoring

Abstract

The rapid growth of IoT devices and their integration in various ecosystems creates a need for high scalability, efficient storage, and fast processing of large amounts of data in real-time. Due to the high frequency of arrival and variety of data formats from sensors, the storage architecture must be flexible and provide high performance for write operations. This creates a challenge in choosing the optimal database management system that would combine the speed of processing large volumes of data with the ability to execute complex analytical queries on historical data effectively. The article presents the results of an experimental study of the performance of the PostgreSQL relational database and the MongoDB document-oriented database in the context of an IoT-based air quality monitoring system. A series of controlled load tests was conducted with varying data volume (100-10,000 records) and the number of parallel threads (1-50). The results show a nine-fold advantage of MongoDB during write operations (34.56 ms vs. 338.16 ms on average) and an almost two-fold advantage of PostgreSQL during read operations (14.41 ms vs. 25.98 ms). Based on the obtained data, a hybrid approach is proposed that combines MongoDB for operational data collection and PostgreSQL for long-term storage and analytics. Such integration of relational and NoSQL databases is a technically optimal and economically beneficial solution. A cost reduction of 30-40% is confirmed, which makes it attractive for real implementation on an industrial scale