Average Rate Analysis of D2D Coded Caching with Uncoded Prefetching

We consider a cache-aided device-to-device (D2D) network where communication is restricted to be single-hop and the placement phase is orchestrated by a central server. The collection of devices cache the whole file set with uncoded prefetching so that the users' requests can be responded by inter-device multicast communication. Under this setting, prior works have characterized the fundamental performance limits of D2D coded caching either through a deterministic worst-case analysis without considering file popularity or through average-case analysis under a uniform file popularity distribution. However, in the real world, the file popularity may not be a uniform distribution. Investigating the performance of D2D coded caching under heterogeneous file popularity is of great interest. In this paper, for a cache-aided D2D network under an arbitrary popularity distribution, we first derive a new information-theoretic lower bound on the expected transmission rate of any coded caching schemes with uncoded prefetching. We also propose an achievable scheme which preferentially caches popular files and uses minimal but necessary space to cache unpopular files. Then, for a Zipf distribution, we analyze the gap between the achievable rate and the lower bound. Numerical results show the effectiveness of our proposed scheme.