Location and Mobility Aware Resource Management for 5G Cloud Radio Access Networks

Cloud Radio Access Network (C-RAN) has recently gained much attention for 5G and Long Term Evolution Advanced (LTE-A) cellular networks. The recent technology advancement in network virtualization function and software defined radio has enabled virtualization of Baseband Units (BBU) and sharing of underlying general purpose processing infrastructure. All these advancements have made C-RAN feasible and practical. This paper proposes new algorithms for clustering towers based on location and for packing BBU clusters based on the prediction of mobility and traffic patterns and analyzes their complexities. Unlike existing C-RAN studies that compared their performance with that of traditional distributed RAN methods, we evaluate and compare the performance with that of existing C-RAN strategies, based on real cellular tower maps provided by the FCC. The proposed combined clustering and packing algorithms have achieved up to 34.8% better QoS while using only up to 5.8% most towers and 7.4% most hosts than other methods. We believe that the proposed combination of clustering and packing algorithms for C-RAN would be significant for the success of emerging 5G networks.