Delay analysis and buffer management for random access in cognitive radio networks

In this paper, we consider a cognitive radio network in which multiple Secondary Users (SUs) contend to access primary network´s channels with a random access scheme. Our goal is to analyze SUs´ queuing delay performance in terms of mean queue lengths and find a minimum buffer space for which the overflow probability is less than a desired threshold. In general, the considered network can be modeled as a multidimensional Markov chain. However, the enormous state space makes the numerical analysis intractable. Nevertheless, the state space can be reduced to a two-dimensional Markov chain in the symmetric channel condition. By this approach, the optimal contention probability that minimizes mean queue lengths is obtained. Besides, we find the minimum buffer space to meet the desired overflow probability. Average queue lengths from simulations are compared with expected queue lengths obtained by analysis. Simulation results show that these two quantities are close to each other for different packet arrival rates and we can meet the overflow probability condition by the proposed buffer space.