Optical-Wireless Cooperative Sleep Control with Traffic Prediction for Energy Efficient Radio Access Network

Abstract

While 6G necessitates more stringent performance requirements, efforts towards carbon neutrality are prompting a focus on energy-saving in networks. However, the power consumption of distributed units (DUs) in radio access network (RAN) is expected to significantly increase due to the rise in computational load for antenna signals processing resulting from rising traffic volumes. Thus, we propose an energy-saving method for RAN that sleeps unused DUs and optical transceivers using traffic prediction with machine learning. Our evaluation showed that the power consumption reduction effect can be increased more than 20% compared to implementing only DU sleep.