面向高并发的云平台计算资源分配策略设计  

Design of high concurrency cloud platform computing resource allocation strategy

作　　者：刘永清 韩泽华 郭晓娟[2] 张立新 崔粉霞 LIU Yongqing;HAN Zehua;GUO Xiaojuan;ZHANG Lixin;CUI Fenxia(State Grid Information&Telecommunication Group Co.,Ltd.,Beijing 102211,China;Beijing China-Power Information Technology Co.,Ltd.,Beijing 100192,China)

机构地区：[1]国网信息通信产业集团有限公司,北京102211 [2]北京中电普华信息技术有限公司,北京100192

出　　处：《电子设计工程》2024年第18期52-56,共5页Electronic Design Engineering

基　　金：国网信通产业集团两级协同研发项目(5468B1210033)。

摘　　要：针对云计算平台服务器利用效率偏低以及用户请求响应速度较慢的问题,同时兼顾该平台的收益最大化,文中提出一种面向高并发的云平台计算资源分配方法。该方法采用帝国主义竞争算法和K-means算法对云计算平台用户提交的工作负载进行聚类,并使用决策树算法来完成资源的分配,从而提供最佳的资源供应方案。K-means聚类算法可有效提高帝国主义竞争算法中殖民地的初始人口,防止聚类过程中得到局部最优解而忽略了全局最优解,且进一步提升了算法的收敛速度。通过仿真实验证明,所提方法的平均响应时间与CPU利用率均优于K-means聚类及决策树等传统方法,而与K-means相比,该方法的平均响应时间降低了1.017s,CPU利用率则提高了25.38%。Aiming at the problems of low utilization efficiency of cloud computing platform servers and slow response speed of user requests,and considering the maximization of the benefits of cloud computing platform,this paper proposes a highly concurrent cloud platform computing resource allocation method.This method uses the Imperialist Competition Algorithm and K-means clustering algorithm to cluster the workload submitted by users of the cloud computing platform,and then uses the decision tree algorithm to determine the resource allocation,thus providing the best resource supply scheme.K-means clustering algorithm effectively improves the initial population of the colony in the imperialist competition algorithm,prevents the local optimal solution from being obtained while ignoring the global optimal solution in the clustering process,and further improves the convergence speed of the algorithm.The simulation results show that the average response time and CPU utilization of the proposed method are better than the traditional methods such as K-means clustering and decision tree.Compared with K-means,the average response time is reduced by 1.017 s and the CPU utilization is increased by 25.38%.

关 键 词：云计算平台 资源分配 帝国主义竞争算法 K-MEANS 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]