当前位置>主页 > 期刊在线 > 计算机技术 >

计算机技术2018年11期

基于多约束值的动态资源调度策略
郁云
(江苏财会职业学院,江苏 连云港 222061)

摘  要:负载均衡问题本质是资源调度问题,需考虑到资源合理分配、闲置服务集群有效利用、实现负载分摊,最终降低运行节点负载量。本文系统地研究了虚拟网络资源调度模型和相关技术,结合云计算和虚拟化网络特点,构建了虚拟网络调度模型,并实现了一种基于多约束值的动态资源匹配(Dynamic resource matching based on multi constraint value,DRMV)策略。当任务请求到达,DRMV 算法根据任务大小、节点负载量、功率及网络带宽等多个约束值对任务和服务节点进行排序处理。同时,为了降低任务和服务节点匹配时间与成本消耗,DRMV 算法利用服务节点实际负载反馈情况,动态调节系统负载。


关键词:云计算;虚拟化网络;负载均衡;资源动态分配;系统稳定性



中图分类号:TP18         文献标识码:A         文章编号:2096-4706(2018)11-0078-03


Dynamic Resource Matching Based on Multi Constraint Value

YU Yun

(Jiangsu College of Finance & Accounting,Lianyungang 222061,China)

Abstract:The essence of load balancing is resource scheduling problem. In order to achieve load sharing,the rational allocation ofresources should be given to considerate and the idle service clusters should be given to effective utilize. Ultimately,reduce the load of therunning node. Systematically studied the virtual network resource scheduling model and some related technologies,build a virtual networkscheduling model under cloud environment according to the characteristics of cloud computing and virtualization of network,and proposedDRMV(Dynamic resource matching based on multi constraint value,DRMV)strategy. When a request arrived,the DRMV sorted andprocessed the task and service node based on the size of task,the node load,power and network bandwidth and other constraint values.At the same time,in order to cut down the time and cost consumption of matching task and service nodes,DRMV used the service node'sfeedback of actual load,dynamically adjusted the system load.

Keywords:cloud computing;virtualization network;load balancing;resource dynamic allocation;system stability


参考文献:

[1] ANDERSON T,PETERSON L,SHENKER S,et al. Overcomingthe Internet impasse through virtualization [J].Computer,2005,3(4):34-41.

[2] 朱伟,聂秀英. 国外下一代互联网研究试验的新进展 [J].电信网技术,2009(4):27-32.

[3] 宋杰,李甜甜,闫振兴,等. 一种云计算环境下的能效模型和度量方法 [J]. 软件学报,2012,23(2):200-214.

[4] TARI Z,BROBERG J,ZOMAYA A Y,et al. A least flowtimefirst load sharing approach for distributed server farm [J].Journalof Parallel and Distributed Computing,2005,65(7):832-842.

[5] ZAKI MJ,Wei Li,PARTHASARATHY S. Customizeddynamic load balancing for a network of workstations [C].//HighPerformance Distributed Computing,1996. Proceedings of 5th IEEEInternational Symposium on.S.l.:s.n.,1996:282-291.


作者简介:郁云(1987.06-),男,汉族,江苏连云港人,讲师,硕士研究生,研究方向:云计算虚拟网络技术。