Summary
Incumbent coexistence and self‐coexistence are common among wireless regional area networks (WRANs) using TV white space (TVWS) due to their large coverages and opportunistic spectrum access. Among incumbents of WRANs are wireless microphones (WMs) whose frequent ON–OFF activities result in intermittent availability of transmission channels. In this work, WRANs under self‐coexistence in addition to incumbent coexistence are modeled for queueing analysis using two‐dimensional (2D) Markov chains taking into account intermittent channel availability for multichannel operations. The proposed models are created based on the widely used ON–OFF Markov chain model for an individual transmission channel. The Markov chain structures depend on the number of WRANs under self‐coexistence as well as the number of transmission channels. In order to reduce the computational complexity associated with a large number of channels, truncated Markov chain models are proposed so that the scenarios with more channels than WRANs can be analyzed with the same complexity as the scenarios with the same number of channels and WRANs. The matrix geometric method is used to numerically compute steady‐state probabilities of the Markov chain, which are then used to evaluate the mean packet delay performance. These results are then used to quantify the mean packet delay obtained from channel sharing among WRANs under self‐coexistence with multichannel operations. Finally, the impact of varying the number of channels as well as varying the ON‐OFF durations of each channel on the delay performance is observed.