工业大数据时序数据存储安全预警研究

Research on safety early warning of time series data storage for industrial big data

为提高工业大数据背景下工业大数据时序数据(IBDTSD)存储的安全性,研发了IBDTSD存储安全预警系统。该系统分为Hadoop分布式存储子系统与资源管理子系统。资源管理子系统管理IBDTSD系统中的所有部件配置,而Hadoop分布式存储子系统则为IBDTSD存储安全监测的功能实现系统,该系统主要由数据采集与处理模块和异常诊断模块构成,根据最小二乘支持向量机算法,完成IBDTSD存储安全异常检测。结果表明:文中设计的系统进行IBDTSD存储安全预警的均方根误差和平均相对误差都小于其他两种系统,说明文中设计的系统预警误差小、泛化能力高;且当时间序列高于30后,系统的预警检测值与实际值趋于相同,说明时序数据越大,时序数据存储安全检测越精确。同时经过多次实验,确认系统的安全预警时延均值为0.02s,预警效率较高。

In order to improve the security of IBDTSD storage under the background of industrial big data, an industrial big data time series data storage security early warning system is developed. The system is divided into Hadoop distributed storage subsystem and resource management subsystem. The resource management subsystem manages all component configurations in the IBDTSD system, while the Hadoop distributed storage subsystem is a functional implementation system for IBDTSD storage security monitoring. The system consists mainly of data acquisition and processing modules and exception diagnostic modules. The support vector machine algorithm is used to complete the security anomaly detection of industrial big data time series data storage. The results show that the root mean square error and average relative error of the IBDTSD storage security warning are smaller than those of the other two systems. The system has a small early warning error and high generalization ability. When the time series is higher than 30, the warning detection value of the system tends to be the same as the actual value, indicates that the larger the time series data, the more accurate the time series data storage security detection. It shows that the average safety warning delay of the system is 0.02s, and the early warning efficiency is high.