Dominant factors for device-to-device occurrence probabilities in cellular networks

Abstract

Device-to-device (D2D) communication is expected to be part of future cellular networks. It is commonly assumed that D2D links will exist and need to be managed. However, the question with what probability D2D links will occur in a network and by which parameters this occurrence is influenced, has not been posed yet. We assume a certain event flow, consisting of (1) the occurrence of communication requirements (2) device pairing and (3) mode selection, that leads to the existence of a D2D flow. Under this event flow, we are interested in the influence of service penetration among UEs, pairing strategies and mode selection types on the occurrence probabilities of D2D communication. We propose a stochastic geometry based formula framework which captures the mean number of D2D links that occur in a certain area. This framework reveals a saturation effect of D2D flows and an interesting interplay among the device pairing and mode selection step. We show that any existing network that uses the considered action flow can be described with this framework and further, by simulation, that it can be used to correctly predict D2D occurrence probabilities.