Maximizing resource utilization in video streaming systems.

摘要

Video streaming has recently grown dramatically in popularity over the Internet, Cable TV, and wire-less networks. Because of the resource demanding nature of video streaming applications, maximizing resource utilization in any video streaming system is a key factor to increase the scalability and decrease the cost of the system. Resources to utilize include server bandwidth, network bandwidth, battery life in battery operated devices, and processing time in limited processing power devices. In this work, we propose new techniques to maximize the utilization of video-on-demand (VOD) server resources. In addition to that, we propose new framework to maximize the utilization of the network bandwidth in wireless video streaming systems.;Providing video streaming users in a VOD system with expected waiting times enhances their perceived quality-of-service (QoS) and encourages them to wait thereby increasing server utilization by increasing server throughput. In this work, we analyze waiting-time predictability in scalable video streaming. We also propose two prediction schemes and study their effectiveness when applied with various stream merging techniques and scheduling policies. The results demonstrate that the waiting time can be predicted accurately, especially when enhanced cost-based scheduling is applied. The combination of waiting-time prediction and cost-based scheduling leads to outstanding performance benefits. The achieved resource sharing by stream merging depends greatly on how the waiting requests are scheduled for service. Motivated by the development of cost-based scheduling, we investigate its effectiveness in great detail and discuss opportunities for further tunings and enhancements. Additionally, we analyze the effectiveness of incorporating video prediction results into the scheduling decisions. We also study the interaction between scheduling policies and the stream merging techniques and explore new ways for enhancements.;The interest in video surveillance systems has grown dramatically during the last decade. Auto-mated video surveillance (AVS) serves as an efficient approach for the realtime detection of threats and for monitoring their progress. Wireless networks in AVS systems have limited available bandwidth that have to be estimated accurately and distributed efficiently. In this research, we develop two cross-layer optimization frameworks that maximize the bandwidth optimization of 802.11 wireless network. We develop a distortion-based cross-layer optimization framework that manages bandwidth in the wire-less network in such a way that minimizes the overall distortion. We also develop an accuracy-based cross-layer optimization framework in which the overall detection accuracy of the computer vision algorithm(s) running in the system is maximized. Both proposed frameworks manage the application rates and transmission opportunities of various video sources based on the dynamic network conditions to achieve their goals. Each framework utilizes a novel online approach for estimating the effective airtime of the network. Moreover, we propose a bandwidth pruning mechanism that can be used with the accuracy-based framework to achieve any desired tradeoff between detection accuracy and power consumption. We demonstrate the effectiveness of the proposed frameworks, including the effective air-time estimation algorithms and the bandwidth pruning mechanism, through extensive experiments using OPNET.