Green and Cooperative DASH in Wireless D2D Networks

We consider the problem of using Dynamic Adaptive Streaming over HTTP (DASH) to a cluster of co-located mobile devices that have both an expensive unicast cellular interface, as well as a free broadcast Device-to-Device (D2D) interface. Under this environment, the traditional DASH on each device may compete the scarce mobile bandwidth to download the same and redundant video contents. In this paper, we propose a D2D DASH framework to promote cooperation among devices, while enhancing video quality of service. The video content server collects status of each mobile device and determines the appropriate bitrate to it, while the mobile device downloads the video fragments and then shares with other mobile devices by D2D communication. In order to save energy while still maintaining video quality, the cluster of mobile devices are divided into Super Node and Peer Node, and where only Super Nodes fulfill the video downloading via expensive cellular networks and then share with others through WiFi broadcasting. We implement the proposed system on an Android testbed, and illustrate its efficient performance by real implementation. Experimental results demonstrate that the bitrate selection algorithm can reduce energy consumption up to 25 % and meanwhile maintain competitive video quality.