集群负载均衡约束下差别数据分离存储仿真  

Differential Data Separation Storage Simulation under Cluster Load Balancing Constraints

作　　者：徐亦丹[1] 龚文辉[1] XU Yi-dan;GONG Wen-hui(Institute of Technology.East China Jiao Tong University,Nanchang Jiangxi 330100,China)

机构地区：[1]华东交通大学理工学院

出　　处：《计算机仿真》2019年第9期452-455,共4页Computer Simulation

基　　金：江西省高等学校教学改革研究课题(JXJG-17-35-5)

摘　　要：数据分离存储利于高效管理差别数据,当前相关研究成果存在安全性与实时性差的问题,提出集群负载均衡约束下差别数据分离存储方法。将差别数据项对应至二维平面中节点,并于二维平面中检索到食物源,根据属性O-measure度量值确定食物源的异常数据,并将具备属性O-measure值最优食物源及其相对应的适应度函数值保存至异常表。为了有效避免人工蜂群在检索异常数据过程中陷入局部收敛情况,将模拟退火算法引入其中,实现异常数据快速查找与剔除。依据数据安全分析,利用谱分量相似度将差别数据分离,并将属性相同的数据归类存储在同一数据集合。实验结果表明,所提方法数据存储过程中安全性能强,实时性好,且存储效率高。The separate storage of data is conducive to efficient management of differential data.In this article,a separate storage method of differential data under cluster load balancing constraint was proposed.The difference data items were corresponded to the nodes on the two-dimensional plane,and the food source were found from two-dimensional plane.According to O-measure value of attribute,the abnormal data of food source was determined.The optimal food source with O-measure value of attribute and corresponding fitness function value were saved to the abnormal table.In order to effectively avoid the artificial bee colony falling into local convergence in the process of retrieving abnormal data,the simulated annealing algorithm was introduced to achieve the rapid search and elimination of abnormal data.According to data security analysis,the difference data was separated by spectral component similarity.The data with the same attribute were classified and stored in the same data set.Simulation results prove that the proposed method has strong security,good real-time performance and high storage efficiency during the data storage.

关 键 词：集群 负载均衡 差别数据 存储 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]