Dynamic Adaptive Streaming over HTTP and Progressive Download: Comparative Considerations

Video data now accounts for over 50% of Internet traffic, and the widespread adoption of smart devices with high resolution screens has led to an increasing sharing of video content over social networks. The high availability of video content and of devices capable of displaying video data with high quality has facilitated this development. The high variability in available bandwidth to different clients on mobile and fixed networks implies that streams of different quality and thus bit rate should be used for different clients. Moreover, Internet congestion fluctuations suggest that the bandwidth available to each client will vary over time. As much of the Internet video has been based on progressive HTTP based downloading, which in its basic form is unable to adapt to bandwidth changes, the Dynamic Adaptive Streaming over HTTP - DASH - standard was proposed to enable adaptability of bit rate during a streaming session. Furthermore, this adaptation is managed by the clients to avoid server scalability issues. In this paper we consider a range of issues related to the performance of DASH and perform experiments that indicate the benefits associated with it.