Research On Several Problems Of Coded Caching

The coded caching technology is to add local cache for users in the network,so as to convert the network traffic cost during peak hours into a cheap cache cost.This technology is of great significance for reducing the network load and improving the network utilization efficiency during peak hours.In coded caching system,there is a central server,which links to a database storing files and multiple users through a shared channel,and each user has its own local cache.The coded caching scheme is divided into two phases,namely”placement phase”and”delivery phase”.In the placement phase,the central server stores part of the file information in the user's cache in advance according to the cache capacity.In the delivery phase,all users send a request to the central server,and the central server sends encoded data to each user in the form of communication multicast.Then users can decode the requested file information according to the stored content and the delivered data.This paper delves into the coded caching problem in different scenarios,proposes the specific prefetching scheme and transmission scheme from the perspective of achievability,proves the rate achievability.This paper also gives the lower bound of the”memory-rate”trade-off from the perspective of the converse under different scenarios which analyzes the advantages and disadvantages of the proposed schemes.The first scenario is when there are users in the encoding cache network who do not request information or leave the network.In order to solve the unnecessary waste of network transmis-sion load,we propose a novel uncoded prefetching,centralized coded caching solution:Robust Coding Caching Scheme.In the prefetching phase,storage content is added to the cache of K~?users.In the delivery phase,some users do not request file information during peak hours.The remaining K active users request a file from the central server.The robust coding cache scheme can ensure the optimal network transmission performance after users leaving,and the achievability of optimal rate and file decoding correctness are demonstrated by the achievabil-ity proof.Besides,we prove that the performance of robust coded caching can meet the lower bound of”memory-rate”trade-off based on information theory.The second scenario is that each user in the network may have a multiple files request.The number of files requested by different users at the same time is arbitrary.In order to satisfy all the requests from all users at the same time,we propose a novel uncoded prefetching,centralized coded caching solution under K?3:Multi-Requests Coded Caching.In this paper,the specific placement scheme and delivery scheme are given in a multi-requests coded cache system when the number of users is K?3.Then we prove the achievability of the transmission rate and the decodability under arbitrary request types.And the lower bound of the“memory-rate”trade-off for arbitrary number of users K is given from the perspective of the converse,proving the optimality of the proposed scheme.When the number of users is K?4,the transmission rate of the actual coded cache scheme has a certain gap with the lower bound of the theoretical rate.Then we analyze the reason why the theoretical optimal rate is unreachable through some specific examples.