北京邮电大学学报

  • EI核心期刊

北京邮电大学学报 ›› 2012, Vol. 35 ›› Issue (4): 89-93.doi: 10.13190/jbupt.201204.89.jinj

• 研究报告 • 上一篇    下一篇

MapReduce架构的多控制节点改进

金晶1,王妍2,2,李昕3,陈山枝2   

  1. 1北京邮电大学 网络与交换技术国家重点实验室, 2河南省电力公司信阳供电公司 3电信科学技术研究院
  • 收稿日期:2011-09-21 修回日期:2011-11-01 出版日期:2012-08-28 发布日期:2012-07-08
  • 通讯作者: 金晶 E-mail:keyman_king@163.com
  • 作者简介:金晶(1983-),男,博士生,Email:keyman_king@bupt.edu.cn 陈山枝(1969-),男,教授,博士生导师
  • 基金资助:

    国家高新技术研究发展计划项目(2010AA012501);国家重点基础研究发展计划项目(2009CB320505);国家科技重大专项项目(2010ZX03-005-002-02,2011ZX03005-004-02);中央高校基础研究基金项目(2011PTB-00-34,2011PTB-01-01)

A New Multi-Master Framework of MapReduce

JIN Jing1,WANG Yan2,LI Xin1,CHEN Shan-zhi3   

  1. 1State Laboratary of Switching and Networking Technology2Xinyang Power Supply Company of Henan Electronic Power Company 3China Academy of Telecommunication Technology
  • Received:2011-09-21 Revised:2011-11-01 Online:2012-08-28 Published:2012-07-08
  • Supported by:

    ;the Fundamental Research Funds for the Central Universities

摘要:

MapReduce集群系统的单一控制节点结构使其存在着性能瓶颈问题,限制了集群的规模. 为较好地解决系统性能瓶颈问题,提出了MapReduce架构的多控制点改进. 新系统的实现基于Hadoop MapReduce,在对架构进行改进的同时,重新设计了信令系统,并增加了控制节点间的热备份机制. 因此,新架构除了能解决瓶颈问题外,还可以通过控制节点间的热备份策略有效地应对控制节点故障,降低故障对集群性能的影响,保证应用的正常执行. 

关键词: 云计算, MapReduce架构, 多控制节点, 性能瓶颈

Abstract:

In the existing MapReduce framework, the single master is the performance bottleneck. As a result, the size of the cluster is limited. In order to break the bottleneck, a multimaster framework is proposed in this paper. The implementation is based on Apache Hadoop MapReduce. The main framework and signaling system are redesigned to sustain more than one master. Meanwhile a doublemaster hot backup scheme is designed to solve single master failure, so that the new framework is able to reduce the impact of master failure.

Key words: cloud computing, mapreduce framework, multi-master, performance bottleneck

中图分类号: