摘要
针对边缘计算服务器运维难度大、运维成本高等问题,基于对单台服务器动态采集的多维指标,并通过汇聚算法得到业务汇聚指标,通过巡检模型评估需要进行优化的目标服务器,并根据动态调整模型给出的决策方案对服务器实施优化调整,提出一种边缘计算服务器性能评估及优化模型。实验结果表明,本模型可实现边缘计算服务器的业务规划、软件配置与硬件配置等决策预警,可提升26%的服务质量、10%的运维效率,同时能降低3.57%的运维成本。
To solve the problem of difficult operation and high maintenance cost of servers of edge computing,a performance evaluation and optimization model of edge computing servers is proposed based on the multi-dimensional indicators dynamically collected from a single server and the business aggregation indicators obtained by the aggregation algorithm,in which the target servers to be optimized are evaluated by the patrol model,and the servers adjusted and optimized according to the decision-making scheme from the dynamic adjustment.The results show that this model can realize early warnings in decision-making process for business planning and software and hardware deployment of edge computing servers,improving its quality of service by 26%and maintenance efficiency by 10%while reducing its operation cost by 3.57%.
作者
张旭
崔建峰
杨威
ZHANG Xu;CUI Jianfeng;YANG Wei(School of Software Engineering,Xiamen University of Technology,Xiamen 361024,China)
出处
《厦门理工学院学报》
2022年第3期56-63,共8页
Journal of Xiamen University of Technology
基金
厦门市教育科学“十四五”规划2021年度课题项目(21038)。
关键词
边缘计算服务器
性能评估
性能优化
监控预警
edge computing server
performance evaluation
performance optimization
monitoring and early-warning
