| The interest in scalable video streaming has increased dramatically. Unfortunately, the number of video streams that can be supported concurrently is highly constrained by the required real-time and high-rate transfers, which quickly consume server and network resources, including network bandwidth and disk I/O bandwidth. Resource sharing techniques face this challenge by utilizing the multicast facility. The main classes of these techniques are Batching, stream merging, periodic broadcasting, and composite techniques. Resource sharing techniques face this challenge by utilizing the multicast facility.;The decision as to which class and particular technique to apply greatly impacts the overall system performance and the perceived quality-of-service (QoS). With the many available techniques, it is unclear which one is the best to use in a target environment. In addition, the achieved resource sharing depends significantly on the user access patterns as well as the available server, client, and network resources. Unfortunately, the overwhelming majority of prior studies assumed simple workload, in which videos are accessed sequentially from the beginning to the end (this pattern is referred to here as Full-Content Access) and clients have homogeneous resources (particularly available download bandwidth and buffer space).;This thesis provides a detailed analysis of resource sharing techniques, considering both the True Video-on-Demand (TVOD) and the Near Video-on-Demand (NVOD) models and two video workloads: mixed-video and hot-video. Guided by this extensive analysis, this work proposes an efficient Workload-Aware Hybrid Solution (WAHS) that combines the advantages of stream merging and periodic broadcasting. Moreover, we propose a Statistical Cache Management (SCM) approach, which computes periodically video access frequencies and determines the data to be cached based on these statistics.;In addition to its performance effectiveness, it is easy to implement and incurs small overhead as updates are triggered only when the workload varies considerably. In addition, we study how to support heterogeneous receivers while delivering video streams in a client-pull fashion. We propose three solutions to address the variability of the download bandwidth among clients: Simple Hybrid Solution (SHS), Adaptive Hybrid Solution (AHS), and Enhanced Hybrid Solution (EHS). We also study how to address the variations in client bandwidth during a session. In addition, we study the support for the variability in the available buffer space among clients. Furthermore, we study how the waiting playback requests for different videos can be scheduled for service in the heterogeneous environment, capturing the variations in client bandwidth and buffer space. Moreover, we study the impact of selected-content access on streaming servers delivering data in a client-pull fashion using stream merging techniques. We propose several enhancements to reduce server load and improve the client perceived quality-of-service (QoS).;Finally, we investigate utilizing more advanced video coding, such as Layered Video Coding (LVC) and Multiple Description Coding (MDC), with advance stream merging techniques for better serving heterogeneous receivers. |
