Analysis of Computational Complexity and Power Consumption in Cloud Based Heterogeneous RAN

Abstract

In the recent past, we have witnessed steep growth in mobile data consumption. To address the capacity requirements resulting from the huge growth in mobile data traffic, the mobile network operators (MNOs) are adding more base stations and allocating more spectrum layers including outdoor and indoor small cells. Since the capacity requirement of the network varies over time, the scaling up of the network may increase the energy consumption of the Radio Access Network (RAN). Hence, we need to optimize the network to reduce the overall power consumption through Cloud based models, and deployment of power-efficient radio nodes. In this paper, we analyze the network evolution towards Cloud based Radio Access Network (CRAN) for a heterogeneous set of base stations such as those with Macro RRUs, Micro RRUs and Pico radio units. We derive the computational complexity using a flexible and ‘future-proof’ power model and apply it for the network. We also compare the computation complexity for various cases of User Equipment (UE) channel conditions, different sub-components within the given base station type and provide the results. We further use the Bin-Packing algorithm to analyze the number of base station cloud servers needed for this network and the power consumption of the base station cloud. We further evaluate whether the newer cloud servers with higher CPU cores are power efficient for a given load. We observe from the simulations, that the currently available base station cloud servers have more capacity and still are more power efficient than the baseline Compute Node servers used with the earlier power model.