Spectral efficiency optimization in caching enabled ultra-dense small cell networks

Abstract

As a complement to content placement and delivery, content retrieving strategies could guarantee that users are timely served via retrieving the desired content via backhaul. With growing deployment of small cell base stations (BSs), nevertheless, over-retrieving may result in excessive interference as well, which degrades the utilization of available spectrum resources. In this paper, we present an analytical framework to evaluate the spectral efficiency of content caching and retrieving in ultra-dense small cell networks (USCN). Specifically, it is shown that, given the amount of cached content, area spectral efficiency (ASE) of USCN would diminish to be zero with over-deployment of BSs and users if content retrieving is not properly designed. On this account, we propose a probabilistic content retrieving strategy and further study the optimization of the content retrieving probability (CRP). Using the derived near-optimal CRP, ASE could be substantially improved. Remarkably, ASE could converge to be a constant instead of zero in the dense BS regime.