Sleep Mode Management for Energy Savings in Realistic 6G Cell-Free Network Scenarios

Abstract

We propose a novel sleep mode management algorithm for access points in 6 G cell-free networks that is able to respond to both large- and small-timescale traffic load fluctuations, utilises only realistically available information and aims to minimise energy consumption under coverage and throughput performance constraints. The simulation-based algorithm assessment and optimisation of its configuration is conducted for realistic scenarios, characterised by a non-negligible transition time for access points to enter/exit a deep sleep mode, a network deployment using available lampposts in downtown Amsterdam and a spatial user distribution based on demographic data for the same area. The results indicate that energy savings of 17.11 % can be achieved if a single load-agnostic algorithm configuration is to be chosen, or 21.54 % for load-optimised configurations, both relative to a baseline scenario without a deep sleep mode. To illustrate the importance of considering a realistic transition time into/out of deep sleep, these energy savings are found to be as high as 31.23 % and 33.46 %, respectively, if we had unrealistically assumed transitions into/out of deep sleep to be instantaneous, as considered elsewhere.