Efficient resource management in multimedia streaming networks.

摘要

Since multimedia applications are among the most popular applications on the Internet, it is critical to efficiently manage the resources in multimedia streaming systems. In this dissertation, we first focus on the quality assurance issue in video streaming across the best-effort Internet, and then investigate efficient resource management schemes in both wide-area networks and broadband cable networks.; To address the issue of random quality degradation in streaming across the best-effort Internet, we design a practical technique, named staggered two-flow video streaming. We deliver a stored video through two flows in a staggered fashion via a VPN pipe from a central server to a proxy server. One flow containing the essential portion of the video is delivered using a novel controlled TCP (cTCP), and the other flow containing the enhanced portion of the video is transmitted using a rate-controlled RTP/UDP (rUDP). We design several application-aware flow control and adaptation approaches to control bandwidth sharing and interactions among flows. Our experiments and simulations have demonstrated the efficacy of the proposed technique in protecting essential data and significantly reducing the numbers of retransmitted/lost packets and the sizes of video prefixes required.; Although Broadband Cable Networks (BCNs) is the most important form of broadband service in the US, it is still lack of effective approaches to exploit their unique characteristics for efficiently providing Video-On-Demand (VOD) service. We propose an optimal full-sharing scheduling approach to address this issue. Through analysis and simulation, we have shown that our approach has remarkable advantages in minimizing the bandwidth consumption of VOD service on BCNs. Furthermore, we design two adaptation algorithms which not only keep bandwidth consumption minimal but also significantly reduce mean service delays. We further analyze the expected bandwidth and the blocking probability of videos, and design a video assignment mechanism to maximize system profits.; We also present a formal analysis of the upstream scheduling of BCNs, identify important relations among scheduling parameters, and formulate an optimal scheduling problem. We further propose several optimized scheduling algorithms to capture the unique input characteristics in the upstream scheduling and reduce request contentions.