Cloud-Based Spectrum Sharing in Virtual Wireless Networks

This paper studies the network-wide spectrum sharing problem in virtual cellular networks applicable to Cloud Radio Access Network architecture. Virtual wireless networks share network resources such as time-frequency resources on the same physical infrastructure. A critical problem is then the allocation of shared physical resources to the virtual networks so that the utilization of the resources is maximized. We formulate the problem as a linear integer maximization problem, which is shown to be NP-hard. We then develop a polynomial time heuristic algorithm called Non-balancing Spectrum Sharing (NSS), which is guaranteed to achieve a solution whose resource utilization approaches half of that of the optimal solution in the worst-case. Two additional heuristic algorithms are also proposed to improve the worst-case performance of NSS by enabling load balancing among adjacent base stations. We have simulated the proposed algorithms and the optimal algorithm under different network configurations. The simulation results confirm that i) NSS performs remarkably close to the optimal algorithm, and ii) the two other heuristic algorithms outperform NSS, and consequently are even closer to the optimal algorithm in the simulated scenarios.