In this paper, we propose a QoS framework for wireless data applications that accounts for the heavy-tailed nature of Internet traffic. Instead of the conventional wisdom of provisioning more conservatively for such traffic, we investigate scheduling strategies that implicitly penalize the long transactions (e.g., long file transfers) that contribute to the heavy tails, while providing good performance to short transactions (e.g., a short Web page download). The proposed framework is illustrated by the simple model of a bottleneck link, at which transactions (e.g., TCP connections) whose lengths follow a heavy-tailed distribution begin according to a Poisson process. The link scheduler we investigate is the foreground-background (FB) scheduler, which serves, in each transmission slot, the transaction that has received the least service thus far. Comparing the conventional round robin (RR) scheduler with the FB scheduler, we find that the FB scheduler requires much fewer network resources to attain a given QoS. After showing the efficacy of the FB scheduler for a single bottleneck link, we consider a wireless link shared by multiple users, where further gains are obtained by using multiuser diversity, exploiting the differences in the qualities of the channel seen by different users. We generalize our notion of QoS for this scenario, in a manner consistent with the notion of proportional fairness (PF), and investigate scheduling using an appropriate combination of PF across users, and FB across transactions for a given user.


    Access

    Access via TIB

    Check availability in my library

    Order at Subito €


    Export, share and cite



    Title :

    A QoS framework for heavy-tailed traffic over the wireless Internet


    Contributors:
    Shao, Zhenwen (author) / Madhow, U. (author)


    Publication date :

    2002


    Size :

    5 Seiten, 13 Quellen




    Type of media :

    Conference paper


    Type of material :

    Print


    Language :

    English




    Robust Rauch–Tung–Striebel Smoothing Framework for Heavy-Tailed and/or Skew Noises

    Huang, Yulong / Zhang, Yonggang / Zhao, Yuxin et al. | IEEE | 2020


    Non-linear preprocessing of heavy tailed reverberation

    Ricker, D.W. / Cutezo, A.J. | IEEE | 2001


    Angle/Doppler estimation in heavy-tailed clutter backgrounds

    Tsakalides, P. / Raspanti, R. / Nikias, C.L. | IEEE | 1999



    Heavy-tailed distribution of scene duration in VBR video

    Casilari, E. / Reyes, A. / Di´az-Estrella, A. et al. | IET Digital Library Archive | 1999