| In recent years,users' demand for video streaming services is increasing,and the requirements for the quality of services are also increasing.However,traditional video streaming technologies are faced with the dilemma of high cost,difficult deployment,and problems of poor adaptability and low sensitivity under complex network conditions.Therefore,Dynamic Adaptive Streaming over HTTP(DASH)emerges at the right moment,which is easy to deploy and well adaptive.Nevertheless,in order to adapt to different network environment,MPD files used in traditional DASH technology often contain many representation versions.The client needs to manage the video streaming process and select the suitable representation according to different network conditions,which is easily affected by network bandwidth fluctuation,increases the workload of the client,and affects the client's perception of network conditions.Meanwhile,the probability of playback rebuffering and stall results in bad viewing experience to the users.Therefore,in order for the client to respond quickly to the bandwidth changes in the complex network environment,it is a good choice for the server to make precise bitrate decisions and push this information to the client.Server-driven policy not only reduces the requirement on the client and helps adjust the client's media consumption rate,but also realizes smooth video playback with the achievable highest quality level.This thesis presents a Server-side control strategy of adaptive bitrate named Server-Dynamic,where the server decides the representation levels contained in the MPD file,simplifies the representation version.The algorithm is verified under the IPv6 environment with a DASH-based video streaming system.Experimental results show that Server-Dynamic achieves very smooth playback and quick response to network condition changes.Also,Server-Dynamic significantly reduces the average video rebuffering times,and improves the average bitrate under different network environments. |
PDF Full Text Request