Comparison of the Performance Sensitivity to the Primary and Secondary Service Time Distribution in Cognitive Radio Networks

In this paper, the impact of the service time probability distribution of both primary and secondary users on the performance of Cognitive Radio Networks (CRNs) is investigated. In particular, negative exponential, log-normal, and hyper-exponential probability distributions are considered for the service time of primary and/or secondary users. The performance metrics of the considered CRNs are obtained for both different ratios of time-scales between PUs and SUs and different combinations of values of the coefficient of variation of the service times. Furthermore, two evaluation scenarios have been considered: with and without resources for exclusive use of SUs. Numerical results show that system performance is more sensitive to the service time distribution of SUs than to that of PUs. Additionally, it is observed that to achieve the quasi-stationary regime, greater difference of the time scales at which events of PUs and SUs occur is required as the value of the CoV of service times increases.