EECS technical reports

Oskicat catalog record
Conditions of Use

Archive Home Page

Soft ARQ for Streaming Layered Multimedia

Podolsky, Matthew
McCanne, Steven
Vetterli, Martin
Technical Report Identifier: CSD-98-1024
November 1998

Abstract: A growing and important class of traffic in the Internet is so-called "streaming media," in which a server transmits a packetized multimedia signal to a receiver that buffers the packets for playback. This playback buffer, if adequately sized, counteracts the adverse impact of delay and reordering suffered by packets as they traverse the network. If large enough, that buffer can additionally provide adequate delay for the receiver to request that the source retransmit lost packets before their playback deadline expires. We call this framework for retransmitting lost streaming-media packets "Soft ARQ" since it represents a relaxed form of Automatic Repeat reQuest (ARQ). While schemes for streaming media based on Soft ARQ have been previously proposed, no work to date systematically addresses two important questions induced by Soft ARQ: (1) at any given point in time, what is the optimal packet to transmit? And, (2) when and how does a receiver generate feedback to the source? In this paper, we address both of these questions with a framework for streaming media retransmission based on layered media representations, in which a signal is decomposed into a discrete number of layers and each successive layer provides enhanced quality. In our approach, the source chooses between transmitting (1) older but lower-quality information and (2) newer but higher-quality information using a decision process that minimizes the expected signal distortion at the receiver. To this end, we develop a model of our streaming media system based on a binary erasure channel with instantaneous feedback and use Markov-chain analysis to derive the optimal strategy. Based on this analysis, we propose a practical transmission protocol for streaming media that performs close-to-optimal retransmission and can adapt to dynamic network conditions. To demonstrate the efficacy of this protocol, we simulate our system and present results that illustrate significant performance benefits both from layering the media signal and adaptively estimating a retransmission deadline.