EAST plant data storage system based on IoTDB time series database

Abstract

EAST (Experimental Advanced Superconducting Tokamak) has been in operation since 2006. With increasing operation time, the volume of plant data stored in MySQL has steadily grown to several billion rows. The current storage architecture centered around the relational database has shown poor performance when facing massive time series data, so it is crucial to adopt TSDB (Time Series Database) for storing plant data. In initial testing, IoTDB demonstrated a performance advantage of at least 2 times over other TSDBs in terms of both write throughput and large-scale queries for plant data management. However, as the inflexible underlying infrastructure of EAST, the plant database cannot be easily modified directly. To remedy this problem, we propose a MySQL-IoTDB Hierarchical Mechanism (MIHM). Specifically, we utilized a MySQL master–slave to IoTDB cluster design to seamlessly transfer the performance burden from relational database to TSDB IoTDB. Extensive tests on the EAST plant data demonstrate that MIHM-based plant data storage system has increased the write throughput by 20 times and the large-scale data querying speed by 100 times compared to previous systems.