云计算环境下的服务调度和资源调度研究云计算环境下的服务调度和资源调度研究.pdf

优先出版 计 算 机 应 用 研 究 第32 卷 云计算环境下的服务调度和资源调度研究* 谢丽霞，严焱心 ( 中国民航大学 计算机科学与技术学院，天津 300300) 摘 要：云计算中的服务调度与资源调度对云计算的性能有重要影响，在分析现有云计算调度模式的基础上，针对云计 算数据密集与计算密集的特点，提出分层调度策略以实现云计算中的服务与资源调度。分层调度策略对任务进行划分确 定作业优先级，并通过数据局部性和总任务完成率对资源进行分配。数值评价部分应用分层调度与已有调度进行比较。 实验结果表明所采用的调度有效提高了资源利用率，为云服务的进一步研究提供了思路。 关键词：云计算；服务调度；资源分配 中图分类号：TP393 文献标志码：A Analysis of service scheduling and resource allocation based on cloud computing XIE Li-xia, YAN Yan-xin (School of Computer Science Technology, Civil Aviation University of China, Tianjin 300300) Abstract: The service scheduling and resource allocation in cloud environment have a major impact on the whole performance of computing. Following the scheduling model in cloud environment, based on cloud computing with data-intensive and compute-intensive features, a stratified scheduling strategy is proposed to achieve service scheduling and resource allocation in cloud computing. Through this stratified scheduling model, the job priority can be decided by service level and task level. And the resources are allocated by data locality and task completion rate. By comparison with existing scheduling, this stratified scheduling model shows a better performance, and effectively improving the resource utilization. The scheduling model provides effective ideas for further research of cloud service. Key Words: cloud computing; service scheduling; resource allocation 主要的Hadoop 调度算法包括FIFO 算法，公平份额调度算法和 0 引言 计算能力调度算法。FIFO 调度算法根据任务提交时间分配任 云计算是目前最流行的分布式计算模式，它不要求任何用 务，首先到达的作业优先执行，不考虑服务时间，这是一