Font Size: a A A

Investigations On Tfrc Congestion Control

Posted on:2013-08-27Degree:MasterType:Thesis
Country:ChinaCandidate:Z S HuFull Text:PDF
GTID:2248330371989407Subject:Computer application technology
Abstract/Summary:PDF Full Text Request
At present, the streaming media application has the rapid growth.Especially with the rapidly growing popularity of wireless networks, streaming media in wireless networks is growing at a geometric rate of growth. However, in the face of the vigorous development of multimedia applications, Because streaming media has a better claim to quality of service(QoS), TCP as the most widely used end-to-end transport protocol in the Internet, but because of the congestion control strategy which led to its use rate volatility can not meet the requirements of the streaming media applications, rate stability, which led to that the streaming media has the problem of fuzziness、out of sync、not smooth. Today, the UDP protocol-based streaming media occupy a position of initiative, but UDP is the lack of congestion control mechanism, its bandwidth is considered non TCP-friendly, the applications of UDP in streaming media make limited network bandwidth resources to be more scarce and will increase network congestion, and may even lead to network collapse.Thus,UDP is not suitable for streaming media applications.According to the problems faced by the streaming media applications,and TCP-friendly protocol for the skills to meet the requirements of the transmission of streaming media has become the research hosspot.In view of this,there are many TCP-friendly protocols have been proposed,such as the Binomial,TEAR,of TFRC. According to its congestion control mechanism to adjust the network load is divided into windows and two categories based on the rate.The most in-depth study of the rate-based TFRC protocol proposed by Floyd et al. TFRC use packet loss rate to adjust the sending rate, and overcome the defects of traditional TCP congestion control mechanism detected a single packet loss dramatically reduce the sending rate, Thus the TFRC obtain a relatively smooth sending rate changes, In addition, due to the TFRC sending rate formula and the TCP throughput formula is similar, the TCP friendliness of TFRC can get better, and TCP traffic flow a fair share of network bandwidth, Simply put, with the traditional TCP protocol,TFRC is more suitable for streaming media business with the characteristics of the smooth sending rate and so on.TFRC protocol as the mainstream TCP friendly rate control protocol itself also has a variety of limitations, mainly in the following two aspects:1) the traditional TFRC protocol is based on a wired network environment, due to the bit error rate of lower subnet in wired network is very low, once lost considered to be caused by network congestion, rather than the data loss caused by link. In wireless networks, as a result of to multipath propagation, delay spread, fading characteristics, as well as the Doppler effect wireless channel has a high error rate, and the wireless packet loss caused by the high error rate and link can not be ignored. If the default is the congestion packet loss starts congestion avoidance mechanism causing the decline in throughput, which is that we do not want to see.2) Since the TFRC protocol sending rate formula is the rate-based model, the friendly and smooth at the expense of sensitivity to changes in network status. When Network entry and exit congestion status. The response of is TFRC protocol very untimely. These defects are not conducive to the remission of network congestion and the full use of network bandwidth.In allusion to the limitations of TFRC protocol in the wireless network, according to the characteristics of the wireless network, on the basis of the analysis of existing improved algorithm, focusing on how to distinguish origin of packet loss, two new improved algorithm was proposed. Improved algorithm improves its throughput, fairness and utilization under the premise of ensuring friendliness in wireless network. This study results are as follows:First, the wireless link loss has the ocharacteristics of burstiness, random, while the congestion packet loss is a cumulative slow process. According to these features, the improved algorithm based jitter size of packet loss rate to distinguish between packet loss is put forward. When the jitter value of the packet loss rate is greater than the threshold set by the algorithm, the algorithm considers the matter unreliable wireless links, packet loss is a wireless packet loss or packet loss for congestion.When the judgment is a wireless packet loss, the algorithm adopt the response mechanism different from congestion packet loss.Secondly, according to the characteristics of the wireless network link loss as well as the status of intermediate nodes’queue length when the network congestion occurs, TFRC improved algorithm is proposed combined with intermediate nodes to distinguish between packet loss, which according to the jitter of the packet loss rate measured to determine the current network conditions, the improved algorithm reduces congestion control caused by the link transmission errors. The algorithm uses a queue length of the intermediate nodes and the jitter amplitude of packet loss rate to divide the current network status, and then distinguish wireless packet loss and congestion loss by the network status. Finally, take a different rate adjustment mechanism depending on the packet loss reasons. In addition, the algorithm also references the idea of the thought of multiplicative increase multiplicative decrease (for short MIMD), it shall be able to response rapidly when network enter or exit the congestion state.The algorithm is based on the state of the network in which packet loss occurred loss reason distinction, achieved good results. In addition, the operator also uses the the MIMD fast response mechanism, thus increasing the link utilization with real-time response speed.
Keywords/Search Tags:Streaming Media, Wireless network, TCP-friendly congestion control, Packetloss rate
PDF Full Text Request
Related items