An Efficient QoS-Aware Computational Resource Allocation Scheme in C-RAN

Abstract

In this paper, one proposes an approach to optimize the computational resource utilization of baseband unit pools in a Cloud Radio Access Network. The problem of resource allocation is formulated and solved as a constrained nonlinear optimization one, based on the concept of bargaining in cooperative game theory. The goal is to minimize resource usage by on-demand resource allocation, per instantaneous requirements of base stations, whilst taking Quality of Service into account. In the event of a shortage of resources, implying that not all demand can be served at the same time, baseband units are prioritized with a weighting policy. Real-time requirements and the priority of services being run on a baseband unit are the two contributors in calculating the weight in a timeslot. Lower prior baseband units, however, are always guaranteed to receive a minimum of resources to prevent them from crashes. Simulation results in a heterogeneous services environment show a minimum 83% improvement in allocation efficiency, compared to a fixed resource allocation scheme based on peak-hour traffic demands. Results also confirm that, in case of a resource shortage, 100% of the resources are fairly distributed among baseband units, fairness being governed by the weight of the baseband units in the pool.