Abstract
Cloud radio access networks (C-RANs) architecture is a cost-efficient and energy-efficient solution for increasing the capacity of the cellular network. In order to adapt the traffic and reduce the system cost, we propose complementary networking architecture for the C-RAN, which takes advantage of both the C-RAN and traditional base stations (BSs). We propose combining the Neyman-Scott cluster process and the Poisson hole process to model the architecture; then, the interference in the large-scale network with cooperation is well characterized through the Poisson point process approximation. By mixing the two multiple association mechanisms, i.e., the RRH-selection mode and the cooperation mode, we explore the tradeoffs between the area spectrum efficiency (ASE), the mean delay, and the system cost both analytically and numerically. The results reveal that the mean delay is negative correlated to the ASE for small ASE and is reversed for large ASE, while the cost is always positively correlated to the ASE. The cooperation mode increases the useful signal as well as the interference, which becomes dominant when the radius of the RRH cluster is large. The proportion of subframes operating in the two modes can be configured to tradeoff the ASE, the mean delay, and the cost.