Optimal caching for device-to-device content distribution in 5G networks

Content distribution using direct device-to-device D2D communication is a promising approach for optimizing the utilization of air-interface resources in 5G network. Research in this area has indicated, perhaps unexpectedly, that the optimal caching distribution for D2D content does not match the distribution of user demands/requests. Therefore, in order to achieve the best utilization of network resources, the network should bias the distribution of cached content. The optimal caching distribution is dependent on, but not identical to, the content demand distribution. In this paper we model transmitting and receiving devices as member of a homogeneous Poisson Point Process (PPP), and use results from stochastic geometry to derive the probability of successful content delivery in the presence of D2D interference and noise. We formulate an optimization problem, and find the best caching distribution that maximizes the total probability of content delivery, assuming user demands/requests are modeled by a Zipf distribution with exponent γ_r. Our results show that the optimal caching distribution follows a similar trend as the demand distribution, and can also be modeled as a Zipf distribution. The exponent γ_c of the optimal caching distribution is related to γ_r through a simple expression involving the path loss exponent.