摘要
针对当前调度方法对多租户软件调度时,难以保证调度任务服务水平目标、租户间公平性,同时存在任务调度执行效率低的问题,提出基于蚁群优化的云计算模式下多租户软件最优调度方法。多租户软件最优调度机制引人占优资源熵以及占优资源权重的概念使任务调度到多租户需求和供给更为贴合的虚拟机上。在此前提下保障获取资源最少的租户优先获得任务调度,并根据租户的资源使用情况,以最大化全局占优资源熵为目标构建最优调度模型。采用蚁群算法对最优调度模型进行求解,对调度方案可行解添加不同信息素,使得多租户调度服务水平目标、调度执行效率、租户间公平性等综合需求达到最优,即最优解。实验结果表明,所提方法能够保证调度任务服务水平目标、租户间公平性,且有效提高了任务调度执行效率。
This paper presents an optimal scheduling method for multi-tenancy software in cloud computing mode based on ant colony optimization. Firstly, the conception about dominant resource entropy and dominant resource weight was introduced into multi-tenancy software optimal scheduling mechanism so that tasks could be dispatched on the virtual machine that multi-tenancy demand was more appropriate for supply. Based on this premise, the tenant with the least resource was guaranteed to obtain task scheduling preferentially. According to tenant＇s resource usage, the maximized global dominant resource entropy was used to build the optimal scheduling model. Moreover, the ant colony algorithm was used to solve the optimal scheduling model and added different pheromone on the feasible solu- tion of scheduling scheme. Thus, multi-tenancy scheduling service level, scheduling execution efficiency, fairness a- mong tenants and other integrated demands could be optimized, namely the optimal solution. Simulation results show that the proposed method can guarantee the service level of scheduling task and fairness. Meanwhile, it effectively im- proves the efficiency of task scheduling.
作者
杨梅芳
YANG Mei-fang(Henan University Puyang Institute of Engineering,Puyang Henan 457000,Chin)
出处
《计算机仿真》
北大核心
2018年第7期447-451,共5页
Computer Simulation
基金
云计算下高职教育信息化建设研究(2014GGJS-292)
关键词
云计算模型
多租户
软件
最优调度
蚁群算法
Cloud computing model
Multi-tenancy software
Optimal scheduling
Ant colony algorithm
