摘要
高校实验室的建设随着信息化改革的不断深入而进入新的发展阶段,从以往小功能、分散型实验室逐步向资源整合型实验室以及实验平台发展,有着私有云和小型数据中心化的发展倾向,然而功耗和成本成为制约其进一步发展的重要因素。因此,许多优化计算、存储和网络的算法被设计出来以期在节能的同时降低单位计算成本。然而,目前大多优化算法侧重于能效的作业分配,而忽视了高校实验室教学、研究等作业可用性这一基本需求。例如,基于能效分配策略将多个工作负载合并到一台服务器,虽然有助于提升单点效率,但也增加了单点故障的概率,而为了恢复这些工作将花费额外的资源,使得期望优化失效,并且这一故障还有可能导致教学事故的发生,从而直接影响教学效果。文中从可用性角度出发,研究并设计了一种爬山算法来防止故障区域失效,保障可用性的同时降低数据中心的能耗。
With the successive deepening of information reform,the construction of laboratories in colleges has entered a new stage of development,during which some small⁃function and decentralized laboratories in the past are developed gradually into resource⁃integrated laboratories and experimental platforms,which presents the trend of private cloud and small data centralization.However,its power consumption and cost have become important factors that restrict the further development of college laboratories.Therefore,many algorithms to optimize computing,storage and network are designed to reduce the unit computing cost while saving energy.At present,most optimization algorithms,however,focus on assignment of energy and efficiency,but neglect the basic needs of the usability of college laboratory teaching,research and other assignments.For example,multiple workloads are merged into one server on the basis of the assignment strategy of energy and efficiency,which is helpful to improve the single⁃point efficiency,but it increases the probability of single⁃point failure.Therefore,additional resources will be spent to restore these failures,which makes the expected optimization ineffective.In addition,this failure may also lead to teaching accidents,which further affects the teaching effect directly.In the perspective of availability,a mountain climbing algorithm is studied and designed to prevent the fault area failure,ensure availability and reduce the energy consumption of the data center of college laboratory.
作者
汪林云
WANG Linyun(Wenzhou Business College,Wenzhou 325035,China)
出处
《现代电子技术》
2022年第23期182-186,共5页
Modern Electronics Technique
关键词
数据中心
节能算法
实验室建设
失效区域
爬山算法
算法设计
测试结果
data center
energy saving algorithm
laboratory construction
failure zone
mountain climbing algorithm
algorithm design
test result
