Research On Key Technologies Of DASH Dynamic Load Balancing By QoE-driven

In recent years,with the rapid development of the network and the increasing number of streaming media applications,users' requirements for quality of experience(QoE)about mobile applications have gradually increased.Dynamic Adaptive Streaming over HTTP(DASH)can select the streaming media rate according to the network environment and cache status of the client.Therefore,how to better to ensure users' QoE is becoming a research hotspot in the DASH environment.The existing streaming media server load balancing technology does not consider the impact of the DASH client on the users' QoE.Especially in the case of high concurrent user task connections,it is difficult to guarantee the overall QoE of the client user.In view of the above problems,the paper studies the evaluation methods of DASH client QoE model and server-side dynamic load balancing technology optimization strategy:(1)A QoE driven model of DASH client(QDMDC)for DASH client is proposed.Model parameters include:initial buffer delay,pause event,video rate/clarity.By linearly fitting the influencing factors of the QoE evaluation model,the key factors of the evaluation model are determined,and the QoE evaluation model is designed as a driving module to guide the optimization of server-side load scheduling.(2)A QoE-driven DASH dynamic load balancing optimization strategy is proposed.Firstly,by analyzing the impact of the load performance of the streaming media server,you can select the appropriate load performance indicator.While realizing the reasonable utilization of the load resources of the cluster nodes,it can better to solve the problem that the number of user requests in the DASH service increases,and the response time on the server side is too long,the bandwidth is crowded,and the load efficiency is reduced.Through the proposed QoE-driven evaluation model,the server-side load optimization scheduling is guided,and the dynamic load balancing optimization strategy of the design is used to improve the cluster load efficiency,and the goal of comprehensive QoE for DASH service users is improved.Based on QoE-driven evaluation model(QDMDC),the QoE-driven DASH dynamic load balancing optimization strategy can optimize DASH services from user subjective experience and load balancing technology.The experimental results show that the optimization strategy can effectively reduce the initial buffer delay and the number of pause times of the client,ensure the average bitrate and clarity of the video,and improve the effective load balancing efficiency of the server cluster,and improve the users' QoE.