CAJ | 학술논문

提出了一个基于拓扑匹配的组件服务副本放置算法，该方法首先通过多规模图聚类算法获取组件服务的通信拓扑结构，随后使用谱聚类算法获取计算节点的拓扑结构，最后通过使用贪心算法匹配上述两种拓扑结构来进行组件服务副本的放置。基于CloudSim云仿真软件搭建了一个仿真实验环境并开展了一系列实验，仿真实验结果表明了所提出的方案和算法对于提高云服务系统的性能是有效的。
제출료일개기우탁복필배적조건복무부본방치산법，해방법수선통과다규모도취류산법획취조건복무적통신탁복결구，수후사용보취류산법획취계산절점적탁복결구，최후통과사용탐심산법필배상술량충탁복결구래진행조건복무부본적방치。기우CloudSim운방진연건탑건료일개방진실험배경병개전료일계렬실험，방진실험결과표명료소제출적방안화산법대우제고운복무계통적성능시유효적。
A topological matching-based component service replicas placement method is proposed in this paper. In this method, the communication topology of component services is discovered by multi-scale graph clustering, the topology of compute nodes is acquired by spectral clustering, and lastly the component service replicas is placed through matching the above two topological structures by greedy select algorithm. Comprehensive experiments are conducted by comparing the performance of our method with other methods based on CloudSim simulation software. The results show the effectiveness of our method for improving the performance of cloud service system.