云际协作环境下能耗与成本感知的工作流调度方法  

Energy and cost aware method for scheduling workflows in jointcloud cooperation environment

作　　者：文一凭 王志斌 刘建勋 许小龙 康国胜 WEN Yiping;WANG Zhibin;LIU Jianxun;XU Xiaolong;KANG Guosheng(Hunan Provincial Universities Key Laboratory of Knowledge Processing and Networked Manufacture, Hunan University of Science and Technology, Xiangtan 411201, China;School of Computer and Software, Nanjing University of Information Science & Technology, Nanjing 210044, China)

机构地区：[1]湖南科技大学知识处理与网络化制造湖南省普通高校重点实验室,湖南湘潭411201 [2]南京信息工程大学计算机与软件学院,江苏南京210044

出　　处：《计算机集成制造系统》2021年第9期2583-2591,共9页Computer Integrated Manufacturing Systems

基　　金：国家重点研发计划资助项目(2020YFB1707600);国家自然科学基金资助项目(6177219,61873316,61872139);湖南省自然科学基金资助项目(2018JJ2142);湖南省教育厅资助项目(20B222)。

摘　　要：为解决云计算中的效费比、平台锁定和跨域资源管理等问题,目前已提出云际计算模式以便多个云服务实体之间的开放协作并创造云价值。在云际协作环境中,如何高效实现兼顾成本与能耗的云工作流调度是涉及多方云资源管理的一个重要问题。针对该问题,构建了云际协作环境下能耗与成本感知的工作流调度模型,并提出一种相应的云工作流调度方法(ECO)。该方法主要包含任务组选择、虚拟机复用和资源动态管理等策略,可在云际协作环境下调度多个云工作流应用,并在满足截止时间约束的前提下,优化工作流执行成本与能耗。通过仿真实验说明了该算法的有效性。To deal with issues like cost-effectiveness ratio,platform lock-in and cross-realm resource management in cloud computing,the jointcloud computing paradigm was proposed to facilitate the cooperation among multi-cloud service entities and creating cloud value.In jointcloud cooperation environment,effective cloud workflow scheduling considering both energy and cost was an important problem related to the management of multi-cloud resources.To solve such problem,a model for energy and cost aware workflow scheduling in jointcloud cooperation environment was constructed,and a corresponding scheduling method was proposed(named ECO).ECO included policies of task group selection,virtual machine reuse and dynamic resource management,which could schedule multiple workflows with deadline constraints in jointcloud cooperation environment,and reduce the cost and energy involved in the execution of workflows.The result of simulation experiments showed the effectiveness of the proposed method.

关 键 词：云际协作 工作流 能耗 成本 调度 云计算 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]