Research And Implementation Of Performance Metrics Measurement Method For Over-the-top Video Service

In recent years,with the continuous development of network technology and the huge demand for OTT(over-the-top)video,OTT video service has become one of the most popular online services for network users.The development of OTT video service has provided people with a variety of experiences.OTT video service occupies more and more network resources and network bandwidth.However,because the majority of Internet users' bandwidth resources are still limited,in the use of high-definition video and other service,users will be beset with the decrease of the quality of the video by poor network status and low transfer rate.Operators are aware of the importance of improving the quality of experience(QoE),which can improve the user experience,and gain higher user retention rates.Therefore,OTT network operators or video content providers urgently need to establish a set of OTT video service metrics measurement system and video service quality guarantee system.Because of the diversity of influencing factors of video quality and the user experience,there has not formed a unified standard on a comprehensive and effective evaluation of OTT video service performance and the actual situation of the quality at home and abroad.Besides,the traditional video measuring technical literature focused on RTSP,which based on UDP/RTP,for real-time streaming media.On the other hand,the domestic and foreign research mainly focus on the establishment of the evaluation model based on user subjective feeling.These technologies are not applied for objective metrics measurement,which based on HTTP/TCP transmission.Which means the study of the method to measure and obtain performance metrics has the very big development space.In this paper,we research the methods mainly from the OTT video transmission conditions,video attributes and user experience directly.We present an Over-the-top Video metrics measuring system for measuring OTT video transmission and playback quality.Metrics(e.g.buffering ratio,rate of buffering)are calculated by analyzing the HTTP stream and the attributes of the video itself(e.g.bitrate,framerate),which is superior to traditional tools.Besides,a passive packet capture mechanism is provided to minimize user perception.Meanwhile,adaptive system architecture is designed to identify various video operators,matching different HTTP content.This paper firstly introduces the research background and significance.Then analyzes and compares the latest research work and existing product of OTT video service metrics measuring,presenting the main content and research goals.And then introduces technology for researching OTT video service metrics measurement methods.After introducing the commonly used measurement metrics,each measuring method is put forward.After that,we analyze the needs of measurement system and expound the summary of system design,detailed design and code implementation.By deploying to the end user's network nodes or the client itself,system can measure OTT video quality and the user experience effectively and passively.Meanwhile,this paper introduces the adaptive architecture of the system,which is conducive to more video website and more extension of measurement method.Finally,we tests the measuring function through the experiment,compared with actual condition,and verifies the accuracy of the measurement results of system.