Performance Analysis for D2D-Enabled Cellular Networks With Mobile Edge Computing

Abstract

In this paper, we propose an analysis framework for the D2D-enabled cellular networks with mobile edge computing, which is composed of two types of computing nodes, i.e., access point (AP) nodes and device-to-device nodes (DNs). The mobile users (MUs), whose locations are modeled as a Poisson Point Process (PPP) distribution, make an independent selection to offload their computation tasks to either the nearby AP or the D2D nodes according to the assumed task offloading modes. Besides, in order to alleviate the intra-tier interference both in uplink and downlink, a prioritized channel access policy (PCAP) is explored. By using the stochastic geometry and queuing theory, the performance of the proposed system model is evaluated in terms of the total successful edge computing probability, which is influenced by both the communication and computation performance of the two offloading modes. We obtain interesting tradeoffs that depend on the effects of D2D node densities and distance thresholds of the assumed offloading mode selection on the above performance metrics. When compared to the scheme without PCAP, the PCAP can contribute more to the total successful edge computing probability of a MU. The derived results could provide an useful guideline for the provisioning and planning in employing D2D-enabled cellular network with edge computing units.