Аңдатпа
The adoption of Digital Twins (DT) and Industrial Internet of Things (IIoT) systems necessitates efficient database solutions for real-time data ingestion and analytics.This study evaluates the performance of time-series databases, Influx Database (InfluxDB) and Timescale Database (TimescaleDB), alongside Not only Structured Query Language (NoSQL) database Mongo Database (MongoDB).Through comprehensive benchmarking, including write throughput and query latency under simulated IIoT workloads, the study identifies trade-offs between write-intensive and read-intensive operations.The results highlight the suitability of InfluxDB for highfrequency data ingestion and TimescaleDB for complex analytical queries.The findings provide actionable recommendations for database selection in digital twin architectures, offering insights for practitioners in industrial applications.Key features and differences, such as data write/read speed and scalability, are analysed.Special attention was given to load testing using Go language, which allowed running parallel threads and achieving write speeds up to 300,000 records per second in InfluxDB.TimescaleDB showed stable performance when executing complex SQL queries, providing 40 ms per query when sampling 50,000 and 250,000 rows.Examples of using time series databases for storing and processing real-time data from IoT sensors are considered.A brief analysis of the OpenTwins architecture, its databases, and internal components related to database operations has been conducted.It is concluded that the choice of technology should be based on specific requirements for data processing speed, analytics, and longterm storage.