Game Based Wireless Fronthaul C-RAN Baseband Function Splitting and Placement

Abstract

Cloud radio access network (C-RAN) architecture has been a key technology and trend of the evolving fifth generation (5G) cellular networks, because it can achieve efficient cooperative communications and reduce deployment and operational costs. In C-RAN systems, although the fiber fronthaul has been a great candidate, we should continue to research the wireless fronthaul for the cost saving and flexible network deployment. With the data volume growth exponentially, however, there are lots of baseband samples need to be transmitted through fronthaul between remote radio heads (RRHs) and the Base Band Unit (BBU) pool. It has been a key obstacle in the adoption of the C-RAN architecture. There is another problem in CRAN with wireless fronthaul, imperfect channel state information (iCSI) result in the gap of RRH and BBU pool. In this paper, we propose game based C-RAN baseband functions splitting and placement to address the problems. In order to increase applicability and commodity of the models furthest, we focus on two typical C-RAN structures, P2P model and RRH cluster model. Based on the model of P2P structure, we propose Courtnot douploy model with incomplete information to split the baseband functions between RRH/RRH cluster and BBU pool. In the RRH cluster structure model, we propose the Nash bargaining solution to split the baseband functions inside RRH cluster. Through the calculated equilibriums using the proposed game models, we can get the perfect tradeoff between fronthaul cost and CRAN cooperative performance in different deployment scenarios. Simulation results show that effective and efficient splitting and placement schemes can be achieved by using the proposed the Courtnot douploy model and the Nash bargaining solution.