Mobile Cooperative Downloading Mechanisms Based On The Super Network Framework

It has become a main behavior of users to download files because of the burst of various network applications of video and music.Along with the increasing number of wireless devices brought by Internet of Thing and Vehicular Ad-Hoc Networks,the existing mobile network facilities(provided by Internet Service Providers)can hardly meet the increasing download demand of wireless users.On one hand,mobile devices probably move too fast to make stabile Internet connections;on the other hand,the facilities can be probably overwhelmed by the massive download requirements.Cooperative downloading combines the mobile network facilities with self-organized networks,which is an important technique to resolve the contradictory between the download demands of users and the serving capacity of the network facilities.The existing studies on cooperative downloading did not take the characteristics of wireless environments fully into consideration,such as rapid changing of topologies,limited computing resources and unreliable data transmissions,so they cannot apply to mobile environments well.Super network is a new ubiquitous network framework with self-management techniques as a core.It can integrate heterogeneous networks,and improve the adaptability of applications.Hence,studying cooperative downloading based on the super network framework can improve its adaptability in mobile environments.Cooperative downloading involves several problems,such as cooperative cluster organization,file coding methods,schedule methods,incentive and security mechanisms.In this work,we only focus on the first three problems.We first design a cooperative user selection method based on the self-management theories of the super network framework,and then study the file coding and task schedule problems to help cooperative downloading combat with the rapid changing of topologies and unreliable data transmissions,and hence improve the performance of cooperative downloading in wireless environments.The main contents are as follows:Firstly,on the aspect of cooperative cluster organization,we focus on urban scenarios,and propose a cooperative user selection method following the self-management theories of the super network framework.We propose a privacy-preserving driving route comparison method to enhance the sensing ability of nodes.Next,based on the sensing ability of the super network framework,cooperative users are selected with the consideration of the characteristics of urban traffic,including car following and lane changes.The simulation results show that the proposed cooperative nodes selection method can largely improve the stability of the network topology of cooperative downloading systems.Secondly,on the aspect of file coding,it is common to use chunked Network Coding(NC)in wireless cooperative downloading,considering the limited computing resources of wireless devices.However,chunked NC can cause extra communications comparing with dense NC,and the number of the extra communications is relevant to the chunk size.Therefore,setting a suitable chunk size to improve the overall performance of chunked NC is a prerequisite for applying it in wireless cooperative downloading systems.Accordingly,we establish an analysis model using a Markov process with the consideration of the unfixed broadcast nodes of cooperative downloading,and propose chunk size calculation methods to optimize the performance of cooperative downloading system by combining queuing theory with the analysis model.The simulation results show that the proposed calculation methods can largely decrease the completion time of cooperative downloading and the using delay of target files.Thirdly,when the target file of cooperative downloading is big,chunked NC can cause massive extra communications.Therefore,it is important to improve chunked NC to meet the increasing demand of downloading big files.Accordingly,we propose a hierarchical NC method based on local information.By leveraging the information shared in small ranges,we introduce an IDNC layer on chunked NC layer to decrease the communication burden of chunked NC.Furthermore,we propose a selection method based on genetic algorithm to decrease the complexity of chuck set selection at the IDNC layer.Because of the low decoding complexity of IDNC,the proposed method keeps the good properties of chunked NC,such as the independence of chucks and low decoding complexity.The simulation results show that the proposed method can effectively decrease the communication rounds of cooperative downloading and averaged decoding delay comparing with existing work,especially when the target file is big.Lastly,we propose a disorder-avoidance task schedule method for wireless cooperative downloading to solve the block disorder problem caused by the difference of cooperative users.We model task delay using queue theory,and propose a calculation method to solve the block disorder problem in theory.Considering the characteristics of wireless networks,such as unstable cooperative users and dynamic download rates,we implement the disorder-avoidance task schedule method by combining the calculation method with a dynamic task delay prediction mechanism.Simulation results show that the proposed method can decrease the usage delay of blocks caused by the block disorder problem,and cut off the long tail of downloading time caused by the cooperative users with very low download rates.