Interplay of Cache Sizes and File Popularity in Coded Caching
Yuhang Yao, Jinbei Zhang · 2019
Coded caching is a caching technique proposed recently, which can greatly reduce backhaul traffic by exploiting the multicast transmissions over different users, even when they request distinct files. The transmission gain of coded caching depends on the distribution of users' cache size and also the distribution of file popularity. While the impacts of user caches or file popularity has been investigated independently, their joint impact on coded caching is unknown and challenging to be analyzed. In this paper, we propose a scheme considering the cache sizes distribution and file popularity distribution simultaneously. The average transmission rate and the lower bound are both analyzed. The gap between the two rates is at most an O(log K) factor, where K is the number of users. Simulation results show superior performance of the proposed scheme.