The Research Of Personal Cloud Storage Measurement Method Based On QoE

The rapid development of Internet has brought the exponential explosion of information.Traditional way of data storage has been unable to meet the demand of the current network data storage.With cloud storage technology becoming more mature,it solves the key technology of vast users' access to huge amounts of data.Personal cloud storage services,as one of the typical application of cloud storage,not only meet the basic demand of users on data storage,but also provide functions such as file synchronization across the equipment and work synergy.It has become a popular data services.User quality of experience(QoE),as factor that can measure different experience provided by different service,has attracted the attention of more and more users and service providers.Measuring terminal users' personal cloud storage QoE is significant in selecting services and optimizing the services of providers.In view of the research on users' terminal QoE measurement,the main job of this paper is as the follows:Firstly,propose a user-oriented method that passively measures the QoE of personal cloud storage.This paper analyzes the impact of data flow and control flow to terminal users QoE from the perspective of the client,network and architecture.Several indicators are proposed from the end user's perspectives,which can reasonably evaluate the QoE of user-oriented personal cloud storage.According to passive measurement technique,a user-oriented method is proposed to measure QoE of personal cloud storage.The method calculates the QoE indicators in user's operating process through capturing network packets passively in service process from the client,which can reflect the real QoE in users' operating process.Design and realize the measurement tool of personal cloud storage QoE,which is composed of data packet capture module,and network flow analysis module and indicator calculation module.In data packet capture module,there are different network communication process on the client-side,for which,a double buffer process capture packet method based on Winpcap filtering mechanism is put forward.In network flow analysis module,by analyzing the characteristics difference of data flow and control flow,this paper puts forward a network stream clustering method that can optimizes initial center of personal cloud storage.In performing a long-term measurement experiment,this paper measures four world wild popular services,namely Baidu Cloud,Kuaipan,and Google Drive and One Drive.By comparing QoE indicators of different services,this paper analyzes the reasons of the difference.The test results provide not only a reliable opinion for the end user in choosing the appropriate service,but also advices for service providers in optimizing its services,which can further prove the validity and necessity of our measurement methods.