Optimizing Tier-Level Content Placement in Heterogeneous Networks

Abstract

Caching popular contents at base stations (BSs) of a heterogeneous cellular network (HCN) reduces latency in content delivery and alleviates congestion in backhaul networks. Despite the existence of many sub-optimal strategies for content placement, the optimal ones for HCNs remain largely unknown and are investigated in this paper. To this end, we adopt the popular HCN model where BSs are modeled as K tiers of independent Poisson point processes (PPPs). Further, the random caching scheme is considered where each of a database of M files with corresponding popularity measures is placed at each BS of a particular tier with a corresponding probability, called placement probability. The probabilities are identical for all BSs in the same tier but vary over tiers, giving the name tier- level content placement. The network performance is measured by hit probability, defined as the probability that a file requested by the typical user is delivered successfully to the user. Consider the case of a uniform received signal-to- interference (SIR) threshold for successful transmissions. The optimal policy is derived in a simple form allowing sequential computation of the optimal placement probabilities. The result shows that the probability for a particular file-and-tier combination is a monotone increasing function of the file's popularity and the tier's density, transmis- sion power and storage capacity. Furthermore, for the general case of non-uniform SIR threshold, the optimization problem is non-convex and a sub-optimal placement policy is designed by approximation. The close- to-optimal policy has a similar structure as that in the previous case.