الفهرس 
AbstractOnly 14 pages are availabe for public view
 |  | from | 154 |  |  |
| | | from | 154 | | |
Abstract
Internet-based video delivery system has already become a major application due to users’ growing demand for video content and continuous growth of wired-wireless network technologies. Recently, with the rapid growth in video delivery systems, people ask for more and more video resources, which cause an overload on network resources. Peer-to-Peer (P2P) network is an important component to implement video delivery systems. Efficient deliver of the video resources in P2P networks has become one of the most critical issues and one of greatest concern to users. Also, how to optimally utilize network resources and to manage the overlay topology are important questions in video delivery. In addition, one major quality of video delivery metric is playout latency that can be considered as one of the key challenges that face many researchers nowadays. This thesis presents an efficient streaming scheme that enables a QoS guaranteed video service over the networks. The P2P video delivery system contains four primary entities: (1) Web Server, (2) Media Server, (3) Tracker Server, and (4) Peers. Initially, web server informs peers with available channel lists. Media server has the original video content. Tracker server is responsible for directing video requesters to either media server or other parent peers. Peers are either super peers or ordinary peers. Super peers serve as a chunk contributor while ordinary peers serve as chunk requesters. In fact, three major components are designed and implemented for our proposed scheme: (1) Overlay distribution to organize contributing peers into hybrid mesh-tree topology, (2) End-to-End video content delivery from super peers to orinary peers using chunk selection scheme, and (3) Dynamically pre-fetching related videos according to peer’s available uplink bandwidth. The thesis proposes a dynamic pre-fetching scheme for P2P video delivery system based on the analysis of the user’s behavior. The probability that users continuously use the downlink bandwidth continuously is low, and a large part of the downlink bandwidth remains wasted during a session. Therefore, it is effective to deliver highly recommended videos and cache them into a service buffer prior to user requests. Our work is particularly suitable for the delay-sensitive applications, since such applications considers continuous media streams and guarantees the services with low delay. The simulation results validate that the proposed scheme can efficiently utilize network bandwidth, minimize server load, and initial playout latency. Furthermore, the demand for video delivery over wireless networks has been gradually increasing over the years. Video delivery is gaining popularity among mobile users. The latest mobile devices, such as smartphones and tablets, are equipped with multiple wireless network interfaces. They provide free wireless connectivity to the end users, offering an easy and viable access to a network and its services. Handover latency in wireless networks is vital to some time-sensitive applications, such as VoD, VoIP, interactive games. Mobile WiMAX handover mechanisms suffer from several defects, particularly related to wastage of channel resources, handover latencies, and loss of data.
This thesis presents a mobility scheme and a performance analysis WiMAX. The performance has been analyzed by studying the impact of different scanning schemes. The scanning could be light or dense according to mobility attributes such as scanning threshold, scan duration, scan iterations, and interleaving interval. The mobility scheme aims to investigate the effects of node mobility while delivering a video stream, on the performance of the mobile WiMAX network. The performance is evaluated by measuring several parameters like handover latency, data dropped, throughput, traffic received, and video application end-to-end delay. The simulation results validate that dense scanning scheme significantly reduces overall handover latency and packet end-to-end delay but in contrast reduces application throughput.