面向云数据中心的成本感知容错算法  

Cost-aware Fault-tolerant Algorithm for Cloud Data Centers

作　　者：宋彩利[1] 梁斌 李皎[1] SONG Caili;LIANG Bin;LI Jiao(School of Computer Science,Xi’ an Shiyou University,Xi’ an,Shaanxi 710065,China)

机构地区：[1]西安石油大学计算机学院,陕西西安710065

出　　处：《西安石油大学学报（自然科学版）》2021年第5期129-135,共7页Journal of Xi’an Shiyou University（Natural Science Edition）

基　　金：国家自然科学基金青年科学项目“考虑信任度和权值的区块链数据通信性能优化研究”(61802301);陕西省自然科学基础研究计划一般项目(2019JQ-056);西安市科技计划项目(2020KJRC0101)。

摘　　要：针对云数据中心重调度算法大多是以牺牲云任务的完成时间或者增加云服务提供商的损失为代价的问题,系统分析了虚拟机故障所带来的性能损失和云数据中心容错策略的重调度过程。参考了Amazon云数据中心的真实定价机制,并对Amazon的多个云数据中心的虚拟机价格进行分析。同时,对失效的云任务根据截止时间进行了分类。利用不同配置虚拟机的定价差异提出了面向云数据中心的成本感知容错算法(Cost-aware Fault-tolerant Algorithm for Cloud Data Centers,CAFT)。该调度算法通过提高云任务的故障修复率从而降低云服务提供商的损失。实验仿真结果显示,相比IRW、RI和DTRDT算法,所提算法的故障修复率明显提升,同时,云服务提供商的损失平均降低了1.3%、13%和0.5%。Most of the rescheduling algorithms for current cloud data centers come at the cost of sacrificing the completion time of cloud tasks or increasing the loss of cloud service providers.This paper systematically analyzes the performance loss caused by virtual machine failure and the rescheduling process of cloud data center fault tolerance strategy.The prices of virtual machines in Amazon's several cloud data centers are analyzed referring to the real pricing mechanism of the Amazon cloud data center.At the same time,the failed cloud tasks are classified according to the deadline.Afterwards,the cost-aware fault-tolerant algorithm for Cloud Data Centers(CAFT)was proposed by using the difference in pricing of virtual machines with different configurations.This scheduling strategy reduces the loss of cloud service providers by increasing the failure repair rate of cloud tasks.Experimental simulation result shows that,compared with the IRW,RI and DTRDT algorithms,the fault repair rate of the strategy proposed in this paper is significantly improved.At the same time,the loss of cloud service providers is decreased by 1.3%,13%and 0.5%on average.

关 键 词：云数据中心 成本感知 云任务映射 容错算法 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]