A Research On Techniques Of Coded Caching In Wireless Network

Along with the rapid development of smart devices,the demand for resource in the network will rapidly increase,which bring out higher demand for the throughput of the network.At the same time,The network traffic is of great variety since the number of user and the request of the user are changing in the network.Efficiently utilize the network resource,such as bandwidth,and reduce the congestion in peak-traffic hour become an urgent problem.One of the most powerful solutions is taking advantage of the memories distributed across the network and duplicating some of the content during the off-peak hour,which means we can copy some popular content at the memory nearby or even at the end-users.In the peak-traffic hour,users can look up the resource in the memory nearby before receiving it from the network.This content duplication is called caching,which can balance the network traffic between peak and off-peak hour and reduce the congestion.Traditional model to use cache is only caching parts of the file locally,where no coding strategy involved.When the total quantity of the file is much larger than the local cache size,the gain which is called local caching gain is insignificant.Maddah-Ali and Niesen creatively come up with the coded caching and their caching problem consists of two distinct phases called placement phase and delivery phase.In the placement phase,we precisely arranging the cache content.Thus,we can take advantage of the broadcast nature of the wireless channel and broadcast message to different users.Using the message received from shared link and the cache content,users can recover their requested file.We can not only get the local caching gain,but also achieve a global caching gain which derives from jointly designing both the placement phase and delivery phase.Maddah-Ali and Niesen's solution deals with the model that is symmetry,for example the file size and popularity,users' cache size and distortion requirements and the channel condition mush be same.This is to rigorous and unpractical.Thus,more and more researchers focus on the asymmetry model.Asymmetry makes the caching problem more complicate but more practical.In this paper,we investigate the coded caching problem in two different asymmetry models.Firstly,we consider coded caching problem when users have different cache size.We provide the optimal strategy to utilize user's cache in two user scenario when only one user has cache while another does not have a cache,and come up with an optimal coded caching scheme when both use have a cache and the server has any number of file using memory sharing and the former conclusion of the symmetry problem.We use the technic called computer-aided proof while proving the optimality of the scheme.The second asymmetry model we discus,users have heterogeneous distortion requirement.Since users may use difference device(like the laptop or smart phone)or have different priority(for example,some users have high fidelity requirement when they watching videos,while others many pursue the fluency),they may have different distortion requirements.We come up with an optimal caching scheme in this model when there are two users with same cache and any number of file.In the last,we reconsider the case where both cache size and distortion requirement are different and give out a caching scheme which is better than Yang and Deniz's.This improvement comes from the coded content in the users' cache.