Future Mobile Networks Optimisation using Cloud-RAN Architecture

This thesis investigates the future of the Cloud-Radio Access Network (C-RAN) architecture and its applicability in future mobile networks. Thus, it analyses if C-RAN deployment is worth the trouble of replacing the RAN equipment of a whole mobile network. C-RAN is an evolving mobile network architecture that aims at centralising and virtualising the radio processing functions in three processing units connected by a crosshaul network. Throughout this thesis, the C-RAN architecture is investigated for different functional split options, various crosshaul network opportunities and crosshaul energy consumption. Starting out with a state-of-the-art overview of the different functional split opportunities and quantifying their capabilities according to their location in the radio processing protocol stack. The crosshaul network must comply with heavy requirements to the bitrate and latency and thus different deployment options are analysed including opportunities for sharing the crosshaul network with other traffic sources. Depending on the functional split chosen, the crosshaul bitrate is fixed or varying with user load. A crosshaul bitrate that varies will not occupy the same amount of transport resources all the time. This creates a statistical multiplexing gain enabling capacity to be shared between different sources. However, crosshaul link sharing will add additional latency to the time sensitive transmission, which limits the crosshaul range. Further, this thesis will investigate if existing links can carry crosshaul traffic for improved sustainability. Energy consumption is an important player in the design of future mobile networks, and C-RAN is envisioned to reduce energy consumption in the radio processing units. This work investigates the energy consumption in the crosshaul network by presenting energy consumption models, designed and tested by theoretical calculations. The models show that for different functional splits, the energy consumption in the crosshaul network can reduce 28% considering 4G, but in 5G, it can save up to 99% when different splits are chosen. In conclusion, C-RAN will optimise future mobile networks, but at the cost of user latency and limited opportunities for crosshaul network deployment as well as transport network energy consumption. Thus, C-RAN is not optimal for extreme latency sensitive scenarios. However, depending on the functional split chosen, the C-RAN architecture will bring benefits such as enhanced interference management, mobility management and resource utilisation. Hence, different combinations of functional splits and functional placements will be applicable to different areas depending on cell density and existing crosshaul transport opportunities.