Trade-Off in Energy Efficiency and Response Performance in Green CRNs: Perspective of Two Sleep Timers and Bursty Traffic

The spectrum utilization problem in cognitive radio networks (CRNs) has been addressed in the existing literature from different viewpoints. However, the growing number of wireless Internet of Things (IoT) devices and the need for continuous connectivity are putting a strain on energy resources, leading to challenges in the sustainability of CRNs. Meanwhile, the explosion of the IoT introduces significant amounts of time-sensitive data with burstiness, posing challenges for designing and operating CRNs. To better trade off energy efficiency and response performance, we propose a novel energy conservation scheme subject to a periodic sleep mechanism and two sleep timers, and we introduce a Markovian arrival process to characterize the bursty arrival pattern of user packets in green CRNs. We develop a four-dimensional continuous-time Markov chain to evaluate the proposed energy conservation scheme. Using the matrix-geometric method, we obtain performance indexes, including the mean delay of packets and the energy conservation level of base stations. We provide experimental results to show how sleep parameters and correlation coefficients affect performance indexes, and we compare the proposed energy conservation scheme with the case where one timer is used. Based on the profit function, the relationship between Nash equilibrium and socially optimum spectrum access strategies is investigated, and a reasonable spectrum pricing policy is presented.