| Summary: | Driven by the widespread use of smartphones and the release of a wide range of online packet data services, an unprecedented growth in the mobile data usage has been observed over the last decade. Network operators recently realised that the traditional approach of deploying more macrocells could not cope with this continuous growth in mobile data traffic and if no actions are taken, the energy demand to run the networks, which are able to support such traffic volumes risks to become unmanageable. In this context, comprehensive investigations of different cellular network deployments, and various algorithms have been evaluated and compared against each other in this thesis, to determine the best deployment options which are able to deliver the required capacity at a minimum level of energy consumption. A new scalable base station power consumption model was proposed and a joint evaluation framework for the relative improvements in throughput, energy consumption,and energy efficiency is adopted to avoid the inherent ambiguity of using only the bit/J energy efficiency metric. This framework was applied to many cellular network cases studies including macro only, small cell only and heterogeneous networks to show that pure small cell deployments outperform the macro and heterogeneous networks in terms of the energy consumption even if the backhaul power consumption is included in the analysis. Interestingly, picocell only deployments can attain up to 3 times increase in the throughput and 2.27 times reduction in the energy consumed when compared with macro only RANs at high target capacities, while it offers 2 times more throughput and reduces the energy consumption by 12% when compared with the macro/pico HetNet deployments. Further investigations have focused on improving the macrocell RAN by adding more sectors and more antennas. Importantly, the results have shown that adding small cells to the macrocell RAN is more energy efficient than adding more sectors even if adaptive sectorisation techniques are employed. While dimensioning the network by using MIMO base stations results in less consumed energy than using SISO base stations. The impact of traffic offloading to small cell, sleep mode, and inter-cell interference coordination techniques on the throughput and energy consumption in dense heterogeneous network deployments have been investigated. Significant improvements in the throughput and energy efficiency in bit/J were observed. However, a decrease in the energy consumption is obtained only in heterogeneous networks with small cells deployed to service clusters of users. Finally, the same framework is used to evaluate the throughput and energy consumption of massive MIMO deployments to show the superiority of massive MIMOs versus macrocell RANs, small cell deployments and heterogeneous networks in terms of achieving the target capacity with a minimum level of energy consumption. 1.6 times reduction in the energy consumption is achieved by massive MIMOs when compared with picocell only RAN at the same target capacity and when the backhaul power consumption is included in the analysis.
|
|---|
