Dynamic peer buffer adjustment to improve service availability on peer-to-peer on-demand streaming networks

In general, on-demand video services enable clients to watch videos from beginning to end. As long as clients are able to buffer the initial part of the video they are watching, on-demand service can provide access to the video to the next clients who request to watch it. Therefore, the key challenge is how to keep the initial part of a video in a peer’s buffer for as long as possible, and thus maximize the service availability of a video for stream relay. In our previous research work, we proposed a novel caching scheme for peer-to-peer on-demand streaming, called Dynamic Buffering. The dynamic buffering relies on the feature of Multiple Description Coding to gradually reduce the number of cached descriptions held in a peer’s buffers, once the buffer is full. In this paper, we proposed three description dropping policies for dynamic buffering, called sequence dropping, m-dropping, and binary dropping. In addition, mathematical formulas of the reduced number of buffer adjustments of descriptions and the reduction of the average number of selectable descriptions for m-dropping and binary dropping by factors of the number of receiving descriptions (n) and the number of discarded descriptions (m) are established. Experimental results showed that the m-dropping, $\emph{m}=\left\lceil {\frac{n} {2}} \right\rceil$ generally outperformed m-dropping, $\emph{m}=2$ and binary dropping in terms of service availability. Even though the accumulated reduction of buffer adjustments for m-dropping policies was less than that for binary dropping, the average number of selectable descriptions for m-dropping was much greater than that for binary dropping. Furthermore, Compared with the sequence dropping, the m-dropping, $\emph{m}=\left\lceil {\frac{n} {2}}\right\rceil$, would have much less number of buffer adjustments with little difference of the number of selectable descriptions.