Research And Aplication Of Video Streaming Service Performance Evaluation Method On Cloud Platforms

With the development of cloud technology and the popularization of video applications,more and more cloud platforms on the market are starting to provide video services.Therefore,the evaluation of video service performance on cloud platforms has received extensive attention.However,the innovation points of the state-of-the-art performance evaluation methods are focusing on unstructured data and adding complex workloads related to machine learning.The researches are relatively vacancy aimed at performance evaluation for video streaming service on the cloud platforms.So,it is of great value to make researches on how to evaluate the performance of video streaming service on the cloud platforms.This paper designs the performance evaluation method of video streaming media service on platforms and applies this performance evaluation method to the reality.First of all,this paper makes an analysis on the architectures of video services on cloud platforms,including live video,video on demand and video transcoding.Based on the characteristics of the service architectures,workloads are designed and corresponding metrics are selected in the performance evaluation method.Besides,this paper detects whether there is a degradation trend of video service performance by modified Cox-Stuart.Then,this paper implements a video streaming media service performance evaluation system which provides a visual platform to conduct the experiment.The system includes three main functional modules:data acquisition,data storage and data analysis.The system collects data producing in the experiment and stores them in database.Finally,the performance evaluation method of video streaming service on cloud platform has been applied to Ali Cloud,QCloud and LeCloud to carry out experiments and collect data in order to make an analysis and comparison among three cloud platforms.As the same time,the degradation trend of video service performance has been detected in this paper.The experiments shows that the performance evaluation method designed in this paper can be well applied to the video services provided by cloud platforms.It can not only quantify the performance,but also detect the degradation trend of video service which is of great value in practical application.