摘要
在强干扰环境下,多源数据的异构性和分散性使得不同来源的数据在格式、协议和语义上不兼容,迫使各节点为应对局部干扰独立备份数据,造成邻近区域的数据重复存储,容灾备份与恢复的效率降低。提出一种强干扰环境下多源数据容灾备份与快速恢复算法,采用K-means聚类对异构数据进行分簇处理,通过轮廓系数法确定最优簇数,迭代更新聚类中心实现数据降维。根据聚类后的多源数据设计双重冗余消除机制:空间维度通过关联矩阵识别空间相近数据并消冗,时间维度基于线性回归模型过滤时序冗余。完成双重消冗后,基于簇头的能量状态对数据展开容灾备份。最后,结合多源数据在时间和空间上的连续性特征,建立联合图域模型,通过模型中数据间的关联性设计迭代恢复策略,从而实现多源数据的高效容灾恢复。实验结果表明:在强干扰环境测试中,所提算法展现出卓越性能,动态编码切换延迟稳定维持在20~30 ms,机会捕获备份吞吐量持续保持7~8 GB/s的高位,回溯精度损失率始终低于0.5%,显著优于对比算法。
In a strong interference environment,the heterogeneity and dispersion of multi-source data make data from different sources incompatible in format,protocol,and semantics,forcing each node to independently backup data to cope with local interference,resulting in duplicate storage in adjacent areas and reducing the efficiency of disaster recovery backup and restoration.To address this issue,a multi-source data disaster recovery backup and fast restoration algorithm was proposed in a strong interference environment.K-means clustering was used to cluster heterogeneous data,and the optimal number of clusters was determined by the contour coefficient method.The clustering centers were iteratively updated to achieve data dimensionality reduction.A dual redundancy elimination mechanism was designed based on clustered multi-source data:the spatial dimension identifies spatially similar data through an association matrix and eliminates redundancy,while the temporal dimension filters temporal redundancy based on a linear regression model.After completing dual redundancy elimination,disaster recovery backup of data was performed based on the energy status of cluster heads.Finally,based on the continuity characteristics of multi-source data in time and space,a joint graph domain model was established.Through the correlation between data in the model,iterative recovery strategies were designed to achieve efficient disaster recovery of multi-source data.Experimental tests showed that the proposed algorithm exhibits excellent performance in strong interference environment tests:the dynamic encoding switching delay remains stable at 20~30 ms,the opportunity capture backup throughput maintains a high level of 7~8 GB/s,and the backtracking accuracy loss rate remains below 0.5%,significantly outperforming the comparative algorithms.
作者
严峥晖
卢南方
甘盛霖
张月
易晓峰
YAN Zhenghui;LU Nanfang;GAN Shenglin;ZHANG Yue;YI Xiaofeng(Department of Intelligent Manufacturing and Equipment,Guizhou Vocational Technology College of Electronics&Information,Kaili 556000,China;Department of Intelligent Manufacturing and Equipment,GuiZhou Vocational Technology College of Electronics&Information,Kaili 556000,China;CETC Big Data Research Institute Co.,Ltd.,Guiyang 550022,China)
出处
《国外电子测量技术》
2025年第6期21-28,共8页
Foreign Electronic Measurement Technology
基金
中央引导地方科技发展资金(黔科合中引地[2024]009)。
关键词
强干扰环境
多源数据
容灾备份
快速恢复
strong interference environment
multi-source data
disaster recovery backup
fast recovery
作者简介
严峥晖,硕士,教授。E-mail:YanZhenghui1977@163.com;通信作者:卢南方,硕士,讲师。E-mail:lunanfang1991@163.com。
