Research On Edge Caching And Migration Method For Video Resource In Mobile Learning

With the rapid increase of mobile smart terminals and Io T devices,various services and applications based on mobile devices have become an important part of human life.Especially in mobile learning,the COVID-19 pandemic leads to explosive growth of the demand for mobile learning services.Video services account for the largest proportion in mobile learning applications.At present,the mainstream video service architecture adopts “cloud-end” architecture.Generally,there is a large geographic separation between the users and the cloud server,which results in more routing nodes and higher network latency for end-to-end communication between the users and the cloud.In addition,the limited bandwidth limits the performance of application services.Mobile Edge Cloud(MEC)has been proposed to tackle the above problems in cloud computing architecture by deploying small cloud computing infrastructure at the edge of the network to make application service closer to users.Aiming at mobile learning application scenario,this thesis proposes a video resource caching method based on MEC,which uses Fluency-oriented Video Edge Caching Method(FVECM)as video edge caching service instance deployed on MEC Nodes to improve Qo S(Quality of Service)and Trajectoryprediction-based Edge Service Migration method(TPESM)as service migration method to guarantee the accessibility of FVECM while users are moving between the service areas of different MEC nodes.The main research contents of this thesis are as follows.(1)In order to solve the video playback lagging problem in mobile learning scenarios,this thesis proposes a video resource edge caching method called FVECM.First,the Date-type-based Bandwidth Predicting Method(DBPM)uses historical bandwidth data to predict future bandwidth data.Second,based on the predicted bandwidth data and users’ course arrangement,the Fluency-oriented Video Resource Offloading Strategy(FVROS)uses static resource offloading and real-time dynamic forwarding to reduce the impact of local network congestion on video loading and improve the fluency of video playback.Simulation experiments show that the method proposed in this thesis can effectively reduce the impact of network congestion on smoothness and fluency of playback and significantly improve the quality of service(Qo S)of mobile learning applications.(2)In order to solve the problem of guaranteeing cache service accessibility when users move between MEC service areas,this thesis proposes an edge service migration method based on trajectory prediction called TPESM.First,the Realtime Trajectory prediction based on Markov Process(MP-RTP)is proposed to extract the patterns of users’ life and travel habits from their historical trajectory data and formally constructs a mathematical model based on Markov process for real-time trajectory prediction.Second,the Real-time Migration Target Selection(RMTS)is proposed to select the best migration targets and pre-migrates edge services based on the predicted user movement trajectory and the geographic distribution of MEC nodes and the state of network links between MEC.Simulation experiments show that TPESM effectively reduces the service interruption time when users move between different edge service areas compared with traditional service migration methods.The proposed edge caching and migration technique for video resources in mobile learning uses partial offloading to decrease the usage of MEC storage resources while increasing the fluency of video playback and the accuracy of migration targets selection by prediecting user trajectory,thus improving the quality of service of mobile learning applications.In addition,the technique proposed in this thesis can also be applied to other application scenarios with high requirements for smooth video playback,such as short video services,VOD(Video On Demand)services and so on.