Impact of Buffer Size on Proactive Spectrum Handoff Delay in Cognitive Radio Networks

Abstract

Spectrum handoff (SH) is a vital process to guarantee seamless and effective services of secondary users (SUs) in cognitive radio networks (CRNs). SH delay has a negative impact on the performance of SUs. For simplicity, the PRP M/G/1 queuing model is used in literature to evaluate the SH delay parameters of CUs in a CRN. However, the design of an infinite buffer size queue in a real-time tele-traffic system is not feasible. We present pre-emptive resume priority (PRP) M/G/1/K queuing model comprising of three priority queues: primary user (PU) queue for higher priority PUs, interrupted user (IU) queue for moderate priority interrupted SUs and SU queue for lower priority newly arrived SUs, to derive the SH performance metrcs such as blocking probability and cumulative handoff delay (CHD) of SUs. This chapter analyses the impact of buffer length (K) on blocking probability and CHD for various proactive SH schemes: non-switching, switching and random SH schemes in CRNs. We present and summarise the detailed comparison of results for blocking probability and CHD in terms of PUs’ arrival rate and mobility parameter of spectrum holes for different K under PRP M/G/1/K queuing network model. Results show that the blocking probability decreases and the CHD increases with increasing value of K. For an optimal value of K, the proposed model offers similar performance to the PRP M/G/1 queuing network model.