Research On Multi-layer Coded Caching

In the past decade,with the rapid development of smart devices and web applications,the data traffic of vedio-on-demand increases at an incredible speed and becomes a nain burden of the network.It is an effective way to deploy storage in networks to relieve the network load.Content caching has a long history in reliving the network load by storing the files requested by the users.These caching schemes are all interested in how to allocate the user cache effectively to maximize the local caching gain.However,this caching gain is considerable as long as the size of user cache is order of the size of file set.A new caching scheme called coded caching has get rid of this constraint by creating coded multicast oppoertunities among users,which can derive the global caching gain.Meanwhile,this caching gain can be of great use in large-scale content delivey network.In this article,we reasearched on the application scenarios of coded caching and specific caching scheme.We first introduced the centralized coded caching scheme and decentralized coded caching scheme based on single layer content delivery network.Besides,we also investigated the influence of nonuniform user demands on the performance of coded caching scheme.We found out that the coded caching scheme of different file popularities is asymptotic optimal with the caching scheme under uniform user demands.This will be of great importance to application in more practical scenario.We also discussed the online coded caching scheme,which is a good expansion on caching scheme in practical scenario.Meanwhile,we also discussed the multi-layer coded caching scheme.Specifically,we investigated on the transmission rate analysis of coded caching under two-layer cache scenario and discussed the lower bound and upper bound of coded caching scheme.Finanlly,we generalized the multi-layer coded caching scheme to a network with any middle layer caches.We also proposed a coded caching scheme applied to nonunifrom file popularity.Surprisingly,we found out that the transmission rate of server stays stably as the number of users in the network increases,which is very meaningful in large-scale content delivery networks...