Secure Probabilistic Caching for Stochastic Multi-User Multi-Relay Networks

In this paper, we study a stochastic multi-user multi-relay network, where the data transmission can be overheard by multiple eavesdroppers. The users, relays and eavesdroppers are assumed to arrive at the network subject to homogeneous poisson point process (PPP). Cache equipped at the relays and users can pre-store part of file contents, which can help enhance the transmission security. Hence, it is of vital importance to design an effective secure probabilistic caching strategy to ensure the security of content transmission. To this end, we first consider several cache-aided transmission modes of self-fetch, D2D-transmission, relay-transmission, and source-transmission. We then study the network secrecy performance by analyzing the expression of secure cache throughput and an analytical lower bound. In order to maximize the secure cache throughput, we further optimize the probabilistic caching strategy by using a heuristic algorithm. Numerical results are provided to demonstrate that the proposed strategy outperforms the conventional caching strategies.