Cache-Enabled Device to Device Networks With Contention-Based Multimedia Delivery

This paper studies the performance of large-scale cache-enabled device-to-device (D2D) networks with homogeneous Poisson point process distributed mobile helpers (MHs) and user equipments (UEs). The MHs are assumed to have caching capabilities and able to serve the random requests from the UEs within the collaboration distance of D2D transmission. To avoid collisions among the concurrent transmissions, a contention-based multimedia delivery protocol is proposed, under which a MH is allowed to transmit only if its back-off timer is the smallest among its associated contenders. By applying tools from stochastic geometry, the transmission probability of MHs is derived and analyzed. We then characterize the coverage probability of the randomly requested files and thereby the successful content delivery probability of the cache-enabled D2D network. Based on the obtained results of successful content delivery probability, the optimal probabilistic caching strategy of MHs is investigated. Particularly, under the proposed contention-based multimedia delivery protocol, it is recommended to simply cache the most popular file at each MH to maximize the successful content delivery probability of the cache-enabled D2D network. Simulations are provided to validate our analysis.