• DocumentCode
    3425931
  • Title

    A heuristic approach to performance management in computer networks

  • Author

    Di Stefano, A. ; Lo Bello, L. ; Mirabella, O.

  • Author_Institution
    Inst. of Comput. Sci. & Telecommun., Catania Univ., Italy
  • Volume
    2
  • fYear
    1999
  • fDate
    1999
  • Firstpage
    849
  • Abstract
    Network segmentation, also known as the Network Division Problem, is one of the most effective techniques for computer network performance management. It has been dealt with in recent years by several researchers who have formalised the problem in mathematical terms. The solutions currently available in literature are based on minimisation of certain functionals, called fitness functions which, however, consider a limited number of network parameters, thus privileging only one performance target to the detriment of others. This paper introduces a new fitness function overcoming the limits of existing solutions, as it takes simultaneously into account a wider range of network performance requirements. The paper also provides an optimisation methodology which is then applied to a real operational environment i.e., the network of a large-size industrial plant. Results of a wide series of tests carried out in different typical working scenarios are illustrated and discussed
  • Keywords
    computer network management; performance evaluation; computer networks; fitness function; fitness functions; heuristic approach; mathematical terms; network parameters; network segmentation; optimisation methodology; performance management; Computer network management; Computer science; Electronic mail; Engineering management; Industrial plants; Intelligent networks; Optimization methods; Telecommunication network management; Telecommunication traffic; Testing;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Emerging Technologies and Factory Automation, 1999. Proceedings. ETFA '99. 1999 7th IEEE International Conference on
  • Conference_Location
    Barcelona
  • Print_ISBN
    0-7803-5670-5
  • Type

    conf

  • DOI
    10.1109/ETFA.1999.813081
  • Filename
    813081