FDASH: A Fuzzy-Based MPEG/DASH Adaptation Algorithm

The evolution of broadband connection technologies, along with the explosive growth of mobile handheld devices, has made video content one of the major Internet traffic types. HTTP adaptive streaming (HAS) is one factor assisting the growing popularity of Internet video. One of its standards, the MPEG dynamic adaptive streaming over HTTP (DASH), allows users to access video streams of multiple resolutions available at a central repository. Various adaptation algorithms using the MPEG DASH protocol have been proposed, aiming to adapt the media bit rate delivered at the client to the current network conditions. In this paper, we present the FDASH rate adaptation scheme that employs fuzzy logic to control the buffering time and the video resolution delivered to the client in order to distribute video segments of the best quality, deliver undisrupted video playback, and avoid frequent changes of video resolution. The comparison of FDASH against other rate adaptation schemes shows that the former successfully provides the client with the optimal video rates, avoiding buffer underflows and unnecessary changes of video resolution in case of fluctuations of the available connection throughput.