满足双QoS约束的两阶段云工作流调度算法  

Two-phase cloud workflow scheduling algorithm meeting bi-QoS constraints

作　　者：武丽芬[1] 严学勇[2] WU Li-fen 1, YAN Xue-yong 2(1. Department of Information Technology and Engineering, Jinzhong University, Jinzhong 030600, China;2. Information Support Center, China Unicom Jinzhong Branch, Jinzhong 030600, Chin)

机构地区：[1]晋中学院信息技术与工程系,山西晋中030600 [2]中国联通晋中分公司信息支撑中心,山西晋中030600

出　　处：《计算机工程与设计》2018年第7期1904-1910,共7页Computer Engineering and Design

基　　金：山西省教育科学"十二五"规划课题基金项目(GH15052);晋中学院"1331工程"重点创新团队建设计划基金项目

摘　　要：为提高多QoS约束下的工作流调度成功率,提出满足截止时间和预算双QoS约束的调度算法DBWS。为寻找调度可行解,将工作流调度划分为两个阶段:任务选择和资源选择。任务选择阶段中,利用任务在工作流结构中的自顶向下分级值为任务分配优先级,得到调度初始排序;资源选择阶段中,将截止时间根据任务深度进行子划分,设计时间质量和代价质量两个均衡因素,通过综合考虑两种因素计算资源质量,获得任务执行的最优资源。实验结果表明,该算法在调度成功率与同步满足两种QoS约束上较同类算法更优。For improving the successful rates of workflow scheduling under multi-QoS constraints,a scheduling algorithm meeting bi-QoS with dealine constraint and budget constraint was designed.For searching the feasible scheduling solution,the workflow scheduling was divided into two stages including the tasks selection and the resources selection.In the tasks selection stage,the upward rank value of tasks in workflow structure was used to define the scheduling priority of tasks,which obtained the initial scheduling order.In the resource selection stage,the deadline was divided into the tasks according to the task＇s depth,and two trade-off factors,the time quality and the cost quality of tasks were designed.The proposed algorithm can compute the resource quality through synthetically considering these two factors,which can obtain the optimal resource for the task.Experimental results show that DBWS has better performance than the same type of algorithms in the scheduling successful rate and synchronously meeting bi-QoS constraints.

关 键 词：云计算 工作流调度 多QOS约束 任务分级 资源选择 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]