基于数据挖掘的南网异地容灾数据负载分析及磁盘空间预测  

Data load analysis and disk space prediction of remote disaster recovery data in Southern network based on data mining

作　　者：姜南 梁智勇 吴晏芳 黄浩 魏子力 吴浩珊 JIANG Nan;LIANG Zhi Yong;WU Yan Fang;HUANG Hao;WEI Zi Li;WU Hao Shan(Zhaoqing Power Supply Bureau of Guangdong Power Grid Co.,Ld.,Zhaoqing 526000 Guangdong,China)

机构地区：[1]广东电网有限责任公司肇庆供电局,广东肇庆526000

出　　处：《电力大数据》2020年第12期37-43,共7页Power Systems and Big Data

摘　　要：为保障中国南方电网"6+1"企业级信息系统数据的安全可靠,南方电网公司建设南方电网(肇庆)异地数据级灾备中心,以异地备份的方式作为"6+1"企业级信息系统数据的容灾手段。肇庆异地数据级灾备中心承载着南网总部及南方电网五省的异地备份数据,具有数据量大、覆盖面广、重要性高的特点。但随着信息系统的数据量不断增长,给灾备中心的存储设备带来的压力与日俱增,如何应用大数据分析方法降低设备故障风险,成为运维容灾中心的一个研究方向。本文中采用ARIMA时序模型对南方电网异地数据级灾备中心的存储设备进行磁盘空间预测和负载分析,以实现准确预测未来磁盘容量的变化趋势作为目标,采用主动预测的方式减少或避免因数据量增加导致的系统故障,提高异地备份数据的可用性、可靠性、及时性。In order to ensure the safety and reliability of the "6+1" enterprise level information system data of China Southern Power Grid,China Southern Power Grid Co.,Ltd has built a remote data level disaster recovery center of China Southern Power Grid(Zhaoqing)to use remote backup as the disaster recovery means of "6+1" enterprise level information system data.Zhaoqing remote data level disaster recovery center carries the remote backup data of southern China Power Grid headquarters and five provinces of China Southern Power Grid.It has the characteristics of large amount of data,wide coverage and high importance.However,with the increasing data volume of information system,the pressure on the storage equipment of disaster recovery center is increasing day by day.How to use big data analysis method to reduce the risk of equipment failure has become a research direction of operation and maintenance disaster recovery center.In this paper,ARIMA time series model is used to predict the disk space and load of the storage devices in the remote data level disaster recovery center of China Southern Power Grid.In order to accurately predict the trend of disk capacity in the future,the active prediction method is adopted to reduce or avoid the system failure caused by the increase of data volume,and improve the availability,reliability and timeliness of remote backup data.

关 键 词：负载分析 磁盘空间预测 数据级灾备中心 时序预测模型 虚拟磁带库 

分 类 号：C39[社会学]