A QoE-aware Video Distribution Path Optimization Approach For Real-Time Video Networking

The streaming video has become the largest component of network traffic so far,and with the rapid development of the Internet,it has exploded,and the quality of experience(QoE)has become a common concern of users and network operators.important question.How to improve the quality of the user's viewing video in massive video and meet the viewing experience of the end user is one of the core issues of current video network performance optimization.Aiming at this problem,the paper builds a QoE-aware video distribution strategy based on the Content Delivery Network(CDN),optimizes the backbone network bandwidth resource utilization,and improves the network transmission quality.Further designs the K-means improved algorithm to distribute the strategy.Optimize to effectively reduce the load on the source server and improve the quality of the video viewing experience of the client user.The main work of the thesis is as follows:(1)QoE is a quantitative model of users' subjective recognition of service.The blockiness,ambiguity,motion,frame rate,resolution and video content of video are the main factors that affect users' subjective perception of video.From these factors,a QoE evaluation model based on video content of user interest is constructed,and the QoE performance analysis of the model is carried out.(2)Improve the QoE evaluation model through the video of the user's interest,based on the performance indicators of the model,design a QoE-aware video distribution strategy.Firstly,through the user interest model to analyze the user's interest in video content.Then,the interest bias vector is used to calculate the user's interest proportion of various types of content,and the obtained push content is pre-distributed to the proxy server to reduce the cost and load of the proxy server.Finally,the prefix caching strategy is used to implement prefix caching for pre-distributed content,thus reducing the start-up delay of requests.(3)Considering the problem of proxy server cost,it is very important to select the node location of user proxy server node.Here we use the K-means clustering analysis algorithm to quickly iterate out the best proxy node location for each user,thus realizing the path optimization of video distribution strategy.Finally,the CDNsim simulation system is used to simulate the experiment,which proves that the video distribution strategy proposed in this paper can significantly reduce the response time of the server,increase the hit rate by 6%~10%,and improvethe quality of user experience.