Research And Implementation Of The Cooperative D2D-DASH System

With the recent popularity of mobile device,the video traffic is expanding while the mobile internet access can not meet the growth of demand for media service.Although the traditonal streaming using Real-Time Streaming Protocol(RTSP)has been replaced by HTTP-based streaming using Dynamic Adaptive Streaming over HTTP Protocol(DASH),which can be managed by content delivery networks(CDN)using standard HTTP opti-mization techniques.But the limited capacity of mobile network is still the bottleneck.Recently cooperative DASH has been referred as a new solution.In this thesis,we propose a device-to-device(D2D)cooperative adaptive streaming system,which supports asynchronous downloading and sharing for multi-bitrate videos in the crowdsourced network,to improve the quality of video streaming.We propose a crowdsourced QoE-aware quality adaptation algorithm for DASH and the metrics of QoE consist of network capacity,buffer level and energy.Based on the framework,we implement a D2D cooperative adaptive streaming application,which is developed on a UNIX operating systems and compatible with popular DASH Protocol.Finally we run this application on a Raspberry Pi,which is similar to a mobile device,to evaluate the performance of the application.The evaluation proved the improvement of fairness,efficiency,and stability of our application to other adaptive streaming systems.The thesis organizes as follow.Firstly we introduce the researches on adaptive streaming and bandwidth aggregation.Secondly,we talk about the theroy on crowd-sourced QoE-aware D2D cooperative adaptive streaming.Thirdly,we implement a real system of D2D cooperarive DASH.Finally,we evaluate the performance of this system.