• DocumentCode
    3722465
  • Title

    Performance Metrics of Local Cloud Computing Architectures

  • Author

    Travis Brummett;Pezhman Sheinidashtegol;Debadrita Sarkar;Michael Galloway

  • Author_Institution
    Dept. of Comput. Sci., Western Kentucky Univ., Bowling Green, KY, USA
  • fYear
    2015
  • Firstpage
    25
  • Lastpage
    30
  • Abstract
    Cloud Computing is a rapidly growing branch of distributed computing. There have been many proposed open source cloud architectures. The purpose of this paper is to look at three of the most popular open source architectures: Eucalyptus, OpenStack, and OpenNebula. These three architectures were compared by their ease of installation and performance. The benchmarks performed calculated CPU performance, Network performance, and I/O performance. The CPU performance was measured by a Python program that calculated Pi to a given digit and returned the time taken to complete the task. Our Network performance benchmark made use of two Virtual Machines (VMs) and the Iperf tool to send data over the network. The results were measured by network throughput between VMs. Finally, I/O performance made use of a MySQL table of 100,000 entries and the sysbench tool to benchmark varying amounts of threads by measuring the number of transactions per second. We also ran these benchmarks on a micro instance of Amazon´s Elastic Compute Cloud (EC2) to compare results.
  • Keywords
    "Cloud computing","Virtual machining","Computer architecture","Benchmark testing","IP networks","Computational modeling","Operating systems"
  • Publisher
    ieee
  • Conference_Titel
    Cyber Security and Cloud Computing (CSCloud), 2015 IEEE 2nd International Conference on
  • Type

    conf

  • DOI
    10.1109/CSCloud.2015.61
  • Filename
    7371454