On Improving Video Streaming Efficiency, Fairness, Stability, and Convergence Time Through Client–Server Cooperation

Many studies have shown that the dynamic adaptive streaming over HTTP scheme is limited in achieving efficiency, fairness, and stability. Solutions proposed in the recent literature target these objectives, but tackle them from either client or server side separately. This paper argues the use of a client-server cooperation-based approach to achieve the three objectives. Effectively, information available at the client side, such as buffer occupancy, available throughput, and previous played representation levels, can be used to better control the video streaming efficiency and stability at the client side. On the other hand, information available at the server side, such as the server's shared bandwidth capacity, and the number of connected clients and their corresponding downloading bitrates, can be leveraged to better tune the system fairness at the server side. Furthermore, the envisioned client-server cooperation aims at shortening the convergence time of the different clients to the fair bitrate allocation without affecting the overall system smoothness while increasing or decreasing the bitrates. The proposed approach is evaluated through extensive simulations using the Network Simulator, NS-3. Its performance is compared against that of notable algorithms, such as the FESTIVE [1] and PANDA [2] schemes. The obtained results show that the cooperation between the client and the server defines a promising approach in enhancing the efficiency, the fairness, the stability as well as shortening the convergence time to the fair bandwidth share.