EAST装置技术诊断系统数据高效归档方案设计与实现

Design and Implementation of an Efficient Data Archiving Solution for the EAST Technical Diagnostic System

摘要: EAST技术诊断系统需长期监测装置运行状态，采集的数据类型复杂，涵盖来自磁体、支撑结构、馈线等多个关键部件的温度、流量、电位等物理信号。每一轮EAST实验运行周期长达5~8个月，涉及300多个通道、数十种信号源，采集数据规模可达数10 TB。庞大且持续增长的数据量对系统的归档性能、数据一致性、检索效率以及存储空间管理提出了严峻挑战。研究基于EPICS通过应用Archive Appliance与RDB Archive Engine构建归档系统。实现了对数据的统一管理，增强了历史数据的多用户并发访问能力与系统整体可扩展性。对Archive Appliance和RDB Archive Engine关于PV最高归档速率、PV检索速度、多客户端同时访问历史数据等方面进行性能测试，实验结果表明Archive Appliance相较于RDB Archive Engine更适配EAST主机诊断数据的存储需求。

Abstract: The EAST technical diagnostic system requires long-term monitoring of the device’s operational status. The collected data is complex, encompassing various physical signals such as temperature, flow rate, and electric potential from critical components including the superconducting magnets, support structures, and feeders. Each experimental campaign of EAST lasts from five to eight months, involving over 300 channels and dozens of signal types, with data volumes reaching tens of terabytes. The massive and continuously growing data imposes significant challenges on archiving performance, data consistency, retrieval efficiency, and storage management. This paper presents a data archiving system built on EPICS, utilizing both the Archive Appliance and the RDB Archive Engine. The system achieves unified data management and improves historical data access concurrency and overall scalability. Performance tests were conducted focusing on maximum PV archiving rate, PV retrieval speed, and multi-client access to historical data. The results demonstrate that the Archive Appliance is better suited for the storage requirements of EAST’s host diagnostic data compared to the RDB Archive Engine.