基于加速收敛蜂群算法的资源感知调度器.pdfVIP

CN 43-1258/ TP 计算机工程与科学 第 38 卷第 8 期 20 16 年 8 月 ISSN 1007- 130X Computer Engineering . Science Vo l. 38.No. 8.Aug. 2016 文意编号: 1007- 130X( 20 16)08-1595-07 基于加速收敛蜂群算法的资源感知调度器. 江 涛1 ，袁景凌1 ，陈臭骋1 ，宋华明2 (1.武汉理工大学计算机科学与技术学院，湖北武汉 430070;2. 湖北省咸宁市公安局，湖北咸宁 437099) 摘 要: 为了能有效处理海量数据，进行关联分析、商业预测等，Hadoop 分布式云计算乎台应运而 生。但随着 Hadoop 的广泛应用，其作业调度方面的不足也显现出来，现有的多种作业调度器存在参数设 置复杂、启动时间长等缺陷。借助于人工峰群算法的自组织性强、收敛速度快的优势，设计并实现了能实 时检测 Hadoop 内部资源使用情况的资源感知调皮器. 相比于原有的作31:.调皮器，该调皮器具有参数设 置少、启动速度快等优势。基准测试结果表明，该调皮器在异构集群上，调皮资源密集型作业比原有调度 器快 10%20% 左右. 关键词:云计算;Hadoop; 作业调皮器;人工蜂都;资源感知 中图分类号:TP391 文献标志码:A doi: 10. 3969/j. issn. 1007- 130X. 2016.08.013 Resource-aware scheduler based on bee colony algorithm with fast convergence 2 JIANG Tao ,YUAN J ing-ling ,CHEN Min-cheng , SONG Hua-ming (1. School of Computer Science and Technology .Wuhan University of Technology .Wuhan 430070; 2. Xianning Public Security Bureau of Hubei Province ,Xianning 437099.China) Abstract : 1n order to effectively deal with ma ssive data , correlation analysis , business forecast and so on , the distributed cloud computing platform H adoop has emerged. But with the wide application of Hadoop , the job scheduling problems become prominent. Many existing job schedulers h ave defects such as complexity of parameters setting , long start time and so on. Taking full advantages of the artificial bee colony algorithm in terms of strong self-organizing ability and fast convergence , we design and im- p lement a resource-aware scheduler of H adoop which can detect real- time interna