الفهرس 
IntroductionOnly 14 pages are availabe for public view
 |  | from | 102 |  |  |
| | | from | 102 | | |
Abstract
Distributed caching is a promising technology for modern cellular networks, 5 G and beyond. Its basic idea is to allow cellular users to exchange and store data directly via device-to-device (D2D) communications, by passing the backhaul infrastructure. As such, it provides the network with many benefits, such as higher sum data rates, lower latency, extended coverage, higher energy efficiency, added security, and greater spectral efficiency. It also provides the cellular user with many benefits, such as lower latency, lower cost and higher energy consumption. However, if precautions are not made, the D2D users can cause harmful interference to the cellular users, as both use the same (uplink) frequency. One such precaution is a guard zone—a disk around the base station (BS) where D2D communications are prohibited. This simple and easy to implement approach, which has unfortunately not received enough research attention, is the main theme of the present research work, where two contributions have been made. In the first contribution, a study is made to demonstrate the effectiveness of the guard zone in reducing the interference to cellular users. We leverage for this demonstration the powerful mathematical branch of stochastic geometry. With the assumption that the BSs are deployed as a Poisson point process (PPP), we construct an analytical model to characterize the coverage probability of the cellular user in the presence of D2D nodes. This probability is a principal indicator of the quality of service (QoS) of the cellular user and of enabling successful caching for the D2D user. The numerical results obtained, which are validated by rigorous Monte Carlo simulations, confirm that the guard zone approach is highly effective in protecting the cellular user from potential D2D interference. In the second contribution, the effectiveness of the guard zone is further enhanced by allowing D2D communications for only pairs of devices that are in close proximity. The idea is that the closer the two devices are, the less the transmit power they need to use. Our aim in this contribution was to show that pushing the D2D users away from the BS and simultaneously reducing their transmit power both can bring the interference seen at the BS almost to null. To this end we propose a novel empirical technique which, given a desired level of interference, identifies an upper bound for the distance between the two devices to be paired without exceeding that level. A salient feature of this technique is that it requires for its implementation no software or hardware modification of the device. We assess the performance of the proposed technique using a stochastic geometry model, through which we characterize the
ii
coverage probability of the cellular user. We then validate the analytical findings obtained from the model by intensive Monte Carlo simulation to ensure the correctness of the model and the superior performance of the technique.