• DocumentCode
    1937009
  • Title

    Dynamic Deployment and Management of Elastic Virtual Clusters

  • Author

    Wei, Xiaohui ; Wang, Haibin ; Li, Hongliang ; Zou, Lei

  • Author_Institution
    Coll. of Comput. Sci. & Technol., Jilin Univ., Changchun, China
  • fYear
    2011
  • fDate
    22-23 Aug. 2011
  • Firstpage
    35
  • Lastpage
    41
  • Abstract
    Virtual clusters are the fundamental support of new generation virtual High Performance Computing (HPC) systems. As a high level virtual system component, virtual cluster is designed based on virtual machine and virtual network to provide virtual executing environment for large-scale parallel applications. This paper proposes a novel solution for dynamic deployment and customization of Elastic Virtual Cluster (EVC). In this work, we support customization of virtual clusters, such as OS images, network topologies and cluster software. Virtual clusters are automatically configured with isolated virtual networks (network address, DNS domain, NIS domain, etc.) and software environments. Two data transmission protocols are presented in this paper to accelerate image deployment. In addition, we propose a novel solution that virtual machines (VMs) on the same host share image without the aid of VMM. Experiment results show that virtual clusters can be deployed and managed on distributed physical nodes efficiently and therefore can be used by the upper level resource consumers or applications.
  • Keywords
    cloud computing; grid computing; cloud computing; elastic virtual clusters; grid virtualization; virtual executing environment; virtual high performance computing; virtual machine; virtual network; Computational modeling; Computer architecture; Protocols; Runtime; Software; Virtual machine monitors; Virtual machining; VJM; cloud computing; disk-image deployment; elastic virtual cluste;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Chinagrid Conference (ChinaGrid), 2011 Sixth Annual
  • Conference_Location
    Liaoning
  • Print_ISBN
    978-1-4577-0885-5
  • Type

    conf

  • DOI
    10.1109/ChinaGrid.2011.31
  • Filename
    6051755