Load Adaptive Caching Points for a Content Distribution Network

Abstract

The unprecedented growth in content demand on smartphones has significantly increased the energy consumption of current cellular and backbone networks. Apart from achieving stringent carbon footprint targets, provisioning high data rates to city vehicular users while maintaining quality of service (QoS) remains a serious challenge. In previous work, to support content delivery at high data rates, the number and locations of caching points (CPs) within a content distribution network (CDN) were optimized while reducing the operational energy consumption compared to typical cellular networks. Further reduction in energy consumption may be possible through sleep cycles, which reduces transmission energy consumption. However, sleep cycles degrade the quality of service. Therefore, in this paper, we propose a novel load adaptation technique for a CP which not only enhances content download rate but also reduces transmission energy consumption through random sleep cycles. Unlike a non-load adaptive (deterministic) CP, the performance results reveal that the load adaptive CP achieves considerably lower average piece delay (approximately 60% on average during the day), leveraging the introduction of random sleep cycles to save transmission energy. The proposed CP saves up to 84% transmission energy during off-peak hours and 33% during the whole day while fulfilling content demand in a city vehicular environment.