Coded Caching For Heterogeneous Network

In recent years,the amount of global mobile traffic continues to grow rapidly,and most of the traffic is resource demanding video data.However,due to the video traffic has high temporal variability,network congestion may occur during peak traffic periods,while a lot of bandwidth may be wasted during off-peak hours.In order to reduce network traffic during peak hours and improve network utilization during off-peak hours,caching is proposed to utilize the storage space of local caches across the network and to perform content placement during off-peak hours.Coded caching can take advantage of the coded-multicasting opportunities for users with different demands,resulting in an additional and significant reduction of the peak traffic.However,most existing researches on coded caching are based on a simple homogeneous network,while the network is usual heterogeneous in most realistic scenarios.The heterogeneity includes different file sizes,un-uniform file popularity,different cache sizes and numbers of users connected to each cache.The heterogeneity of the network brings new challenges for the design of coded caching schemes.Thus how to design an effective coded caching scheme in heterogeneous network to reduce the bandwidth becomes a key issue.In this paper,we propose a novel optimization strategy for coded caching that minimizes the worst-case transmission rate upon users requests.In order to effectively solve this problem,we propose a practical algorithm using Lagrange multiplier method and sequential quadratic programming approach.Similarly,considering the other two heterogeneous features in heterogeneous networks,we have re-established the corresponding optimization problem and given the corresponding practical algorithm.And finally we can obtain the caching proportion for different files in different caches.Experiments show that the optimal caching scheme based on the proposed algorithm can effectively reduce the worst-case transmission rate.We also analyze the influences of these heterogeneous conditions on the worst-case transmission rate,and the relationship between the various heterogeneous conditions.Experiments show that when the heterogeneity becomes more obvious,the performance of the coded caching scheme decreases while the gain of transmission rate brought by the proposed scheme becomes more significant.In other words,compared with other schemes,the proposed scheme can effectively reduce the performance degradation caused by heterogeneities.Further,in order to achieve a tradeoff between the worst-case transmission rate and average transmission rate in heterogeneous networks,we propose an improved coded caching scheme.For different file popularities,we partition the files into multiple groups with approximate popularity according to some division method.We allocate different memory to different file groups.On the other hand,for files in the same group,a different caching proportion is also implemented.We model the problem as an optimization problem to minimize the sum of worst-case transmission rate for all groups.By optimizing the allocation of caches between groups of files and the caching proportion for files in the same group.Then we propose a practical algorithm by decomposing the problems to a master problem which updating the allocation of cache space between file groups,while a series of subproblems to determine the caching proportion of files in each file group.We also discuss in detail the method of grouping files and h selecting the number of groups through experiments.At the same time,experiments show that,in the heterogeneous network,the proposed improved coded caching scheme can effectively reduce the worst-case and average transmission rate compared to previous schemes.