(1. 昆明理工大学计算中心, 云南省昆明市 650500; 2. 昆明理工大学电力工程学院, 云南省昆明市 650500; 3. 昆明理工大学云南省计算机重点实验室, 云南省昆明市 650500)
研究了一种基于Lustre分布式并行文件系统的高性能I/O流式数据存储模式,能够满足当前行波暂态采集装置海量暂态数据的高速并发读写访问的需求。通过采用多客户端工作模式,并且将多个暂态数据文件聚合成大文件,可以进一步提高访问带宽和使用效率。实验数据表明该模式具有较高的I/O性能,能够满足目前和未来海量电力数据的高速存储管理需求。
国家自然科学基金资助项目(51267009)。
(1. Computer Centre, Kunming University of Science and Technology, Kunming 650500, China; 2. Faculty of Electric Power Engineering, Kunming University of Science and Technology, Kunming 650500, China; 3. Computer Technology Application Key Lab of Yunnan Province, Kunming University of Science and Technology, Kunming 650500, China)
Massive COMTRADE data storage system should have high I/O performance and good scalability to meet current and future data management needs in an era of big data. This paper is concerned with a high I/O performance stream storage mode based on a Lustre distributed file system. The mode is able to meet the demand for I/O and concurrent access applied in collecting and online analyzing massive COMTRADE data by the gathering device. A multi-clients mode on Lustre and a method of packing multiple COMTRADE files into one big file are adopted to significantly improve I/O scalability. Experimental results show that the mode with its high I/O performance is up to requirement. This work is supported by National Natural Science Foundation of China (No. 51267009).
[1] | 陈永艳,束洪春,董俊,等.基于Lustre文件系统的海量暂存COMTRADE数据分布存储模式[J].电力系统自动化,2015,39(18):75-80. DOI:10.7500/AEPS20141225012. CHEN Yongyan, SHU Hongchun, DONG Jun, et al. An Efficient Distributed Storage Mode About Massive COMTRADE Data Based on Lustre[J]. Automation of Electric Power Systems, 2015, 39(18):75-80. DOI:10.7500/AEPS20141225012. |