流媒体应用及其QoS控制机制的研究

发布时间：2018-05-17 14:52

本文选题：流媒体 + QoS控制　；参考：《中国科学院研究生院（计算技术研究所）》2002年硕士论文

【摘要】： 流媒体是一种在网络上实时流式传输的连续时基媒体用户可以一边下载一边对接收的部分数据进行播放流媒体应用涉及实时技术多媒体技术高吞吐量存储技术高速网络技术通信技术等多种技术适用于远程教育远程医疗视频点播现场直播可视电话视频会议网络互动游戏电子商务多媒体广告实时监控等许多领域本文在Windows Media基础上设计并实现了一个完整的流媒体服务系统主要包括流媒体服务器存储系统播放器节目和用户管理等部分可应用于视频点播现场直播多媒体广告等领域在局域网这类带宽充足的网络环境中流媒体应用的关键是流媒体服务器的性能本文通过对三套视频点播系统进行性能评测确定了流媒体服务器的磁盘外部交换率带宽CPU构成流媒体应用的硬件瓶颈以及流媒体服务器的调度程序构成流媒体应用的软件瓶颈这也是性能优化的主要问题在Internet这类无法可靠保证服务质量Quality of Service的网络环境中拥塞和带宽不足成为流媒体应用的主要瓶颈因此Internet中流媒体应用的关键是QoS控制机制本文设计并实现了一个基于RTP协议和MEPG-4可扩展编码的QoS控制机制这是Internet智能流媒体服务系统研究的一部分该QoS控制机制具有以下三个特点 1保持多媒体流与TCP流的友好性和公平性为防止多媒体流占用过多资源而使传统的TCP流无法正常传输本文限制流媒体的传输速率并采用类似TCP拥塞控制的方式对流媒体进行速率控制这样有效保证了多媒体流和传统TCP流的友好共处公平竞争 2非均匀的QoS控制机制本文利用多媒体数据中各部分重要性的非均匀分布在带宽预分配流调度丢包处理错误处理等方面对不同多媒体流安排不同的QoS控制策略从而实现了带宽处理缓存等资源的合理高效分配 3预防拥塞本文将丢包率持续增大但拥塞尚未最终形成的时间段定义为警戒阶段本文的QoS控制机制在警戒阶段即开始进行流控制从而以较小的代价避免拥塞而不是等拥塞发生后再以较大的代价减轻拥塞和消除拥塞
[Abstract]:Streaming media is a kind of continuous time base media which can transmit real-time streaming over the network. Users can play some of the received data while downloading. The application of streaming media involves real-time technology, multimedia technology, high throughput storage technology. Speed network technology communication technology and other technologies are suitable for distance education tele-medical video on demand live video teleconferencing video conference network interactive games electronic commerce multimedia advertising real-time monitoring and many other fields This paper designs and implements a complete streaming media service system based on Windows Media, which includes streaming media server, storage system, player, program, user management, etc., which can be used in the field of video on demand live broadcast multimedia advertisement and so on The performance of streaming media server is the key to the application of streaming media in a network environment such as LAN with sufficient bandwidth. In this paper, the external switching rate of the streaming media server is determined by evaluating the performance of three VOD systems. Bandwidth CPU constitutes the hardware bottleneck of streaming media application and the scheduler of streaming media server constitutes the software bottleneck of streaming media application, which is also the main problem of performance optimization in Internet, which cannot reliably guarantee the quality of service Quality of Service. Congestion and lack of bandwidth in network environment become the main bottleneck of streaming media application. Therefore, the key of streaming media application in Internet is QoS control mechanism. This paper designs and implements a QoS control mechanism based on RTP protocol and MEPG-4 extensible coding. This is a part of the research of Internet intelligent streaming media service system. The QoS control mechanism has the following three characteristics 1. Keeping the friendliness and fairness of multimedia stream and TCP stream. In order to prevent multimedia stream from taking up too much resources, the traditional TCP stream can not be transmitted normally. This paper limits the transmission rate of streaming media and adopts a similar TCP congestion control method. The rate control of streaming media can effectively guarantee the friendly coexistence and fair competition between multimedia stream and traditional TCP stream. 2 Non-uniform QoS control mechanism this paper uses the non-uniform distribution of the importance of each part of the multimedia data to arrange different QoS control strategies for different multimedia streams from the aspects of bandwidth pre-allocation flow scheduling packet loss handling error handling and so on. The reasonable and efficient allocation of bandwidth processing cache and other resources is realized. (3) Prevention of congestion this paper defines the period in which the packet loss rate increases continuously but the congestion has not been finally formed as the alert phase in which the QoS control mechanism begins to implement flow control at the alert stage so as to avoid congestion at a relatively low cost. Is to wait for congestion to occur and then reduce congestion and eliminate congestion at a higher cost
【学位授予单位】：中国科学院研究生院（计算技术研究所）
【学位级别】：硕士
【学位授予年份】：2002
【分类号】：TP393.02

【相似文献】