Flexible Functional Split Design for Downlink C-RAN With Capacity-Constrained Fronthaul

In cloud radio access networks, functional split refers to a division of signal processing functionalities between the baseband unit (BBU) pool and remote radio heads (RRHs). The functionality of baseband signal precoding can either be performed by the BBU pool or RRHs, which corresponds to different functional splits. The compression-after-precoding (CAP) and data-sharing (DS) strategies are the realizations of these two functional splits. In this paper, we propose a flexible functional split design to enable the dynamic functional configuration of each active RRH to use either CAP or DS strategy. Our goal is to minimize the aggregate power consumption, while taking into account limited fronthaul capacity, fronthaul power consumption, and quality-of-service requirement. We formulate a joint RRH mode (i.e., CAP, DS, and sleep) selection, precoding design, and fronthaul compression problem. The formulated problem is a non-convex quadratically constrained combinatorial optimization problem. Through sequential convex programming and l(1)-norm convex relaxation, the problem is transformed into a sequence of semidefinite programming problems. An efficient algorithm based on the majorization-minimization scheme is developed to solve the problem. Simulations demonstrate the importance of considering the limited fronthaul capacity and the performance improvement of the proposed algorithm compared with the pure CAP and DS strategies.