Optimization of TCP Congestion Control Algorithm in Dynamic Adaptive Streaming over HTTP

With the rapid growth of Internet video traffic, streaming media transmission technologies are also developed rapidly. From the traditional real-time transport protocol (RTP) over user datagram protocol (UDP) to hyper text transfer protocol (HTTP) over transmission control protocol (TCP), major video service providers are constantly developing new media transmission technologies to attract more users. As one of the most popular adaptive streaming media transmission technologies, dynamic adaptive streaming over HTTP (DASH) has many advantages in improving users’ experience. However, due to its fragmented transmission, the resulting ON-OFF transmission mode causes bursts of TCP traffic. The intermittent traffic bursts have an impact on other applications. When multiple users compete for the bandwidth at the same time, the players may estimate the network bandwidth incorrectly. This will cause the players switch among the different video resolutions frequently, and the users experience bad QoE (quality of experience). As the transport layer protocol, TCP congestion control algorithm plays a decisive role in the video transmission efficiency, but the traditional congestion control algorithm is not customized for the DASH streaming media transmission. This paper proposes a TCP congestion control algorithm TCP-HAS which adapts TCP Vegas to the DASH streaming media transmission. More specifically, bandwidth and video’s bitrate are also used to set TCP congestion control parameters in TCP-HAS. Experimental results demonstrate that TCP-HAS can improve network QoS (quality of service) as well as QoE of the users when multiple users share the bottleneck link bandwidth.