• DocumentCode
    3454282
  • Title

    Analytic model for end-to-end packet based higher layer protocols over cell-based networks

  • Author

    Hassan, M. ; Egudo, R. ; Breen, J.W.

  • Author_Institution
    Sch. of Comput., Monash Univ., Churchill, Vic., Australia
  • fYear
    1995
  • fDate
    3-7 Jul 1995
  • Firstpage
    162
  • Lastpage
    166
  • Abstract
    An analytic model for an end-to-end acknowledgement and retransmission-based transport protocol with window flow control over a cell-based switch has been developed; acknowledgements are based on packets of multiple cells. Using this model the effect of packet size, window size, and buffer size on the effective throughput or goodput of on overloaded system, where N such end-to-end sources are multiplexed onto a cell-switch, is investigated. It is shown that, for fixed window size either very small or very large packets will have better goodput. For large window and large packets, increasing the buffer size at the cell-switch will actually decrease the goodput in the overloaded case
  • Keywords
    asynchronous transfer mode; buffer storage; packet switching; queueing theory; telecommunication congestion control; transport protocols; ATM switch; analytic model; buffer size effect; cell-based networks; effective throughput; end-to-end acknowledgement; goodput; overloaded system; packet based higher layer protocols; packet size effect; retransmission-based transport protocol; window flow control; window size effect; Analytical models; Asynchronous transfer mode; Australia; Computer networks; Packet switching; Probability; Propagation losses; Switches; Throughput; Transport protocols;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Networks, 1995. Theme: Electrotechnology 2000: Communications and Networks. [in conjunction with the] International Conference on Information Engineering., Proceedings of IEEE Singapore International
  • Print_ISBN
    0-7803-2579-6
  • Type

    conf

  • DOI
    10.1109/SICON.1995.526038
  • Filename
    526038