基于速率的流媒体网络拥塞控制算法研究

发布时间：2018-05-19 19:55

本文选题：流媒体 + 拥塞控制　；参考：《兰州理工大学》2014年硕士论文

【摘要】：伴随着流媒体的广泛应用和人们对高服务质量的追求,流媒体传输技术成为网络应用的一大热点。流媒体集音频视频及图文于一体,具有数据量大、低时延、低抖动等特性。流媒体数据在网络中传输时常采用的是UDP协议,而非TCP。但是UDP并不提供任何的拥塞控制机制,这使得服务质量不能得到有效保障。并且流媒体传输会消耗大量的网络资源,拥塞控制的缺乏将会影响到网络的正常运行。因此,为流媒体传输引入适当的拥塞控制机制势在必行。实践证明,最有发展前景的拥塞控制算法应该在保证媒体流QoS的同时,又注重与TCP数据流的友好共处。基于以上要求,本文首先分析了拥塞产生的原因、拥塞控制的特点和研究意义,其次对流媒体和实时流式传输作了详细的阐述,接着研究了TCP拥塞控制的基本原理、支持实时传输的RTP和RTCP协议,然后重点研究了基于速率的拥塞控制算法TFRC(TCP-Friendly Rate Control protocol)。针对TFRC算法中缺乏对单向延迟抖动的控制、发送速率变化波动性大、慢启动时间过长等不足,本文对TFRC算法进行改进,提出一种更适合流媒体传输的拥塞控制算法,称之为FS-TFRC (Improved TFRC in respect of Fairness and Smoothness)。利用RTCP反馈信息包获取计算吞吐率的参数RTT和丢包率p,根据吞吐量模型来计算相应的发送速率,从而起到平滑发送速率的作用。特别是在慢启动阶段利用FS-TFRC降低了慢启动时间,将单向传输延迟抖动作为反馈信号来修正FS-TFRC的发送速率,同时采用一种自适应阈值调节方案,根据历史情况动态的调整抖动延迟阂值。最后通过仿真实验,验证了改进算法FS-TFRC在降低时延、减缓抖动和保证平滑性等方面都取得了较好的结果,能够更好地适应流媒体实时传输。
[Abstract]:With the wide application of streaming media and the pursuit of high quality of service, streaming media transmission technology has become a hot spot in network applications. Streaming media with audio, video and text in one, with a large amount of data, low delay, low jitter and other characteristics. Streaming media data is often transmitted in the network using UDP protocol, not TCP. However, UDP does not provide any congestion control mechanism, which makes the quality of service can not be effectively guaranteed. And streaming media transmission will consume a lot of network resources, the lack of congestion control will affect the normal operation of the network. Therefore, it is imperative to introduce appropriate congestion control mechanism for streaming media transmission. It has been proved that the most promising congestion control algorithm should not only guarantee the QoS of media flow, but also pay attention to the friendly coexistence with TCP data stream. Based on the above requirements, this paper first analyzes the causes of congestion, the characteristics and significance of congestion control, and then describes the streaming media and real-time flow transmission in detail, and then studies the basic principle of TCP congestion control. RTP and RTCP protocols are supported for real-time transmission, and then the rate-based congestion control algorithm (TFRC(TCP-Friendly Rate Control protocol) is studied. In view of the lack of one-way delay jitter control in TFRC algorithm, the fluctuation of transmission rate and the long time of slow start, this paper improves the TFRC algorithm and proposes a congestion control algorithm which is more suitable for streaming media transmission. It is called FS-TFRC improved TFRC in respect of Fairness and Smoothnessen. The RTCP feedback packet is used to obtain the parameter RTT and the packet loss rate p.The corresponding transmission rate is calculated according to the throughput model, which plays the role of smoothing the transmission rate. Especially in the slow start stage, FS-TFRC is used to reduce the slow start time, the one-way transmission delay jitter is used as the feedback signal to correct the transmission rate of FS-TFRC, and an adaptive threshold adjustment scheme is adopted at the same time. Adjust the jitter delay threshold dynamically according to the historical situation. Finally, the simulation results show that the improved algorithm FS-TFRC has achieved better results in reducing delay, reducing jitter and ensuring smoothness, and can better adapt to real-time streaming media transmission.
【学位授予单位】：兰州理工大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP393.06

【参考文献】