• DocumentCode
    2229739
  • Title

    Superset: A Non-uniform Replica Placement Strategy towards High-Performance and Cost-Effective Distributed Storage Service

  • Author

    Xiangyu Luo ; Yun Wang ; Zexi Zhang ; Hui Wang

  • Author_Institution
    Key Lab. of Comput. Network & Inf. Integration, MoE Southeast Univ., Nanjing, China
  • fYear
    2013
  • fDate
    13-15 Dec. 2013
  • Firstpage
    139
  • Lastpage
    146
  • Abstract
    Load balance and power proportionality are both important aspects in constructing high-performance and cost-effective distributed storage systems. However, traditional replica placement strategies towards load balance usually produce scattered replica layouts which disable power proportionality, while recent strategies towards power proportionality are typically based on uniform replication which compromises the ability of load balance. In this article, we introduce Superset (an organized non-uniform replica placement strategy) which takes both load balance and power proportionality into consideration. The main idea is to partition the whole system into multiple uniform replication based subsystems with the accommodated file subsets satisfying the ´superset´ condition. We have conducted a series of simulations with real-world distributions of data popularity. Our results show that, compared to state of the art solutions, Superset consumes less energy to fulfill the same performance requirement while offers better performance subject to the same energy consumption constraint.
  • Keywords
    distributed databases; parallel processing; power aware computing; resource allocation; storage management; Superset; accommodated file subsets; cost-effective distributed storage service; cost-effective distributed storage system construction; energy consumption constraint; high-performance distributed storage service; high-performance distributed storage system construction; load balance ability; organized nonuniform replica placement strategy; power proportionality; state of the art solutions; uniform replication based subsystems; Energy consumption; Gears; Joints; Layout; Parallel processing; Rabbits; Throughput;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Advanced Cloud and Big Data (CBD), 2013 International Conference on
  • Conference_Location
    Nanjing
  • Print_ISBN
    978-1-4799-3260-3
  • Type

    conf

  • DOI
    10.1109/CBD.2013.34
  • Filename
    6824586