End-To-End Multimedia Congestion Controlarithmetic Analyse And Research

Because more and more multimedia applications appear constantly, multimedia congestion control mechanism is mainly researched in this article. At present, there are two kinds of multimedia congestion control mechanisms: one kind is based on windows; the other kind is based on rate. Because of multimedia application'specific demand, many multimedia congestion control mechanisms use UDP as transport layer protocol, and multimedia congestion control mechanism should be can coexist with TCP protocol, namely TCP-Friendly. Multimedia transmission have anather character: transmission rate should not be fluctuant frequently, but should be smooth. Now, TCP-Friendly multimedia congestion control mechanism has been hotpoint. Many researches point: multimedia congestion control based on rate is smoother than that based on windows. In many multimedia congestion control mechanisms based on rate, TFRC has been applied widely.The article first introduced TFRC protocol mechanism detaildely and its dynamic performance. On the base, insufficiency of TFRC was put forward: in dynamic net, TFRC cannot adapt well to the change of net circumstance. This induces a lot of date-bages are dropped in a short period and effect video quality badly. By analysed, this insufficiency is mainy because of the TCP throughput model used by TFRC. The TCP throughput model was predigested by the condition, that is p go to zero. Withal, The TCP throughput model did not consider"slow-start"prosecc after overtime. So, when drop rete is small, the throughput formula is accurate relatively, but when drop rate is bad, it cannot accurately reflect the network status any more. At this time, TFRC cannot make right decision. So, TCP throughput formula used by TFRC was improved in the article. In order to validate the effect of improvement, ns-2 and other tools were used, for example, VLC media player and rtptools. They together built a dummy network. At last, compareing the video which passed the dummy network with the priamry video, the improved effect is very clear. The simulations results indicated that TFRC had been improved and so wae the video quality. But, because more parameters was used, compute speed descended and affected transmit rate, it need to be further improved.