考虑故障因素的云服务器的最优策略及性能分析

打开文本图片集

关键词：服务器故障容忍；响应时间-功耗权衡；马尔可夫决策过程；方差分析中图分类号：TP391.9 文献标志码：A 文章编号：1001-3695（2025）12-024-3716-07doi：10.19734/j. issn.1001-3695.2025.05.0161

Optimal strategy and performance analysis of cloud servers considering fault factors

Chen Siyi，Yuan Bo，Kuang Haiyang （SchoolofAutomation&ElectronicInformation，XiangtanUniversity，XiangtanHunan411oo，China）

Abstract：Ina cloud center environment，long-running servers mayexperience failures.Therefore，this study aimed to addressthe server selectionproblem forrandomlyarivingrequests inacloudcenter withpotentialfailureriss，inorderto achievetheoptimaltrade-offbetween expectedresponsetimeand powerconsumption.Theresearch employeda Markovdecisionprocesstoanalyzetheexpectedresponse timeofrequests，proposedaniterativeoptimizationmethodtodeterminethebest serverselectionstrategy thatmetthetargetrequirements，andusedanalysisofvariancetoevaluate theimpactofsystemparametersonthealgorithm’sperformance.Theresultsshowthattheproposed iterativeoptimizationmethodcanefectivelydetermine the best server selectionstrategy，achieve theoptimal balance between expected response timeand powerconsumption. Andtheanalysisofvarianceclarifiesthespecificinfluencelawsofsystemparametersonthealgorithm’sperformance.This studyprovidesanefectivetheoreticalbasisand practical guidanceforcloudserviceproviders toselectserversandoptimize system performance while considering server failures.

Key words： server fault tolerance；response-power tradeoff；Markov decision process；ANOVA

0 引言

近十年来，云计算技术极大地推动了IT业务的发展。（剩余16142字）