• DocumentCode
    3705494
  • Title

    An efficient energy-aware mechanism for virtual machine migration

  • Author

    Leonardo P. Cardoso;Diogo M. F. Mattos;Lyno Henrique G. Ferraz;Otto Carlos M. B. Duarte;Guy Pujolley

  • Author_Institution
    Grupo de Teleinformatica e Automacao - Universidade Federal do Rio de Janeiro (COPPE/UFRJ) Rio de Janeiro, Brazil
  • fYear
    2015
  • Firstpage
    1
  • Lastpage
    6
  • Abstract
    A major concern for cloud infrastructure providers is to decrease both their carbon footprint and their energy consumption. In this paper, we propose an elastic and energy-aware virtual machine migration mechanism that decreases power consumption by turning off physical machines and turns them on according to the demand. We develop heuristics that reallocate virtual machines in a smaller number of physical machines, turns off the idle physical machines and, consequently, saves energy. When the demand increases, the proposal turns on physical machines to supply the required resources. The contributions of the paper are four-fold: i) an efficient and elastic mechanism for energy saving on data centers; ii) a simple formal model for minimization of idle physical machines; iii) an implementation in a real environment, and iv) a simulation of the optimization algorithm for larger scenarios. The mechanism was developed and tested in Future Internet Testbed with Security (FITS). The results show the effectiveness of the mechanism on reallocating virtual machines to decrease power consumption.
  • Keywords
    "Virtual machining","Cloud computing","Resource management","Proposals","Energy consumption","Power demand","Simulated annealing"
  • Publisher
    ieee
  • Conference_Titel
    Global Information Infrastructure and Networking Symposium (GIIS), 2015
  • Type

    conf

  • DOI
    10.1109/GIIS.2015.7347183
  • Filename
    7347183