2nd International IEEE Conference on Communication System Software and Middleware

Research Article

Impact of Retransmission Delays on Multilayer Video Streaming over IEEE 802.1le Wireless Networks

  • @INPROCEEDINGS{10.1109/COMSWA.2007.382410,
        author={Sai Ganesh Sitharaman and Maruthu Anantharaman},
        title={Impact of Retransmission Delays on Multilayer Video Streaming over IEEE 802.1le Wireless Networks},
        proceedings={2nd International IEEE Conference on Communication System Software and Middleware},
        publisher={IEEE},
        proceedings_a={COMSWARE},
        year={2007},
        month={7},
        keywords={802.11  EDCF  FEC/ARQ  Video Streaming},
        doi={10.1109/COMSWA.2007.382410}
    }
    
  • Sai Ganesh Sitharaman
    Maruthu Anantharaman
    Year: 2007
    Impact of Retransmission Delays on Multilayer Video Streaming over IEEE 802.1le Wireless Networks
    COMSWARE
    IEEE
    DOI: 10.1109/COMSWA.2007.382410
Sai Ganesh Sitharaman1,*, Maruthu Anantharaman1,*
  • 1: Wipro Technologies, Bangalore, India.
*Contact email: saiganesh.sitharaman@wipro.com, karthik.ananthu@wipro.com

Abstract

In this paper, we seek to establish probabilistic bounds of retransmission delays for transporting multilayer video frames over IEEE 802.11e QAP/QSTA with enhanced MAC distributed coordination function (EDCF). We consider an end-to-end multilayer video streaming that uses hybrid FEC/ARQ error detection and control. Under multiple priority levels of IEEE 802.11e MAC EDCF, we first establish steady-state collision probabilities and contention resolution delays, given the number of nodes. We introduce a time-varying Rayleigh slow-fading channel error model and studying its effect on MAC EDCF transmissions. For video transmissions, we model the expected waiting time of EDCF MAC video queue using head-of-line (HOL) priority queueing discipline using the MAC delay distribution derived earlier as service distribution. The total MAC EDCF video (base layer) queueing delay is the sum of expected waiting time of high-priority voice frames, service residual of best-effort data and the expected waiting time of video frames at HOL queue. Next, we model video retransmission events at receiver as renewal-reward process of frame(s) identified for retransmission to establish the "spread"-time between successful renewal events. The "spread"-time is indeed the probabilistic retransmission bound that we seek for a single video frame identified for retransmission. We verify our model and analytical bounds using an in-house multimedia mobile communication platform (MMCP), written entirely in software to study the cross-layer interworking between MAC and transport for IEEE 802.11 and 802.11e MAC. MMCP currently supports MPEG4 single-layer and FGS two-layer with concurrent voice and video streaming capabilities. Our model, when combined with a receiver-based channel feedback, can yield a jitter-free, rate-adaptive and guaranteed "base" video quality.