Channel Allocation for Fairness in Opportunistic Spectrum Access Networks

In this paper, we investigate the fairness between two secondary user (SU) groups in an opportunistic spectrum access networks (OSA). First, for a given number of allocated channels the throughput of each SU group is analyzed in terms of its forced termination and blocking probabilities using continuous time Markov model. Based on these calculations, the fairness between two SU groups is discussed for different traffic conditions. The results show that for given datarate and relatively higher service durations of two SU groups, the OSA network is highly unfair in complete spectrum sharing scenario. The channel allocation achieve fairness at a cost of decrease in aggregate throughput. However, the computation of a channel allocation vector is a complex process as it require the knowledge of the termination and blocking probabilities of both SU groups. To overcome this drawback, we propose a suboptimal approach by approximating the behaviour of these probabilities. Numerical results show that the proposed method follows the optimal closely. Simulation results are also provided to support the findings.