• DocumentCode
    2250489
  • Title

    A Novel Blocks Placement Strategy for Hadoop

  • Author

    Xianglong Ye ; Mengxing Huang ; Donghai Zhu ; Peng Xu

  • Author_Institution
    Coll. of Inf. Sci. &Technol., Hainan Univ., Haikou, China
  • fYear
    2012
  • fDate
    May 30 2012-June 1 2012
  • Firstpage
    3
  • Lastpage
    7
  • Abstract
    A novel blocks placement strategy is developed for solving some shortages in recent Hadoop. With the Strategy, the optimal Data nodes, according to the real-time situation of Data nodes, are chosen dynamically to achieve load balancing. Finally, the simulation experiments indicate that our strategy behaves much better than the HDFS blocks placement when the situation is coherent to time.
  • Keywords
    distributed processing; file organisation; public domain software; resource allocation; Google MapReduce programming model; HDFS; Hadoop distributed file system; blocks placement strategy; load balancing; open-source implementation; optimal data nodes; Bandwidth; Fault tolerance; File systems; Load management; Network topology; Real time systems; Switches; HDFS; blocks placement; hadoop; load balance;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Computer and Information Science (ICIS), 2012 IEEE/ACIS 11th International Conference on
  • Conference_Location
    Shanghai
  • Print_ISBN
    978-1-4673-1536-4
  • Type

    conf

  • DOI
    10.1109/ICIS.2012.11
  • Filename
    6211069