Characterization of the Opportunistic Service Time in Cognitive Radio Networks

This paper studies the service time required to transmit a packet in an opportunistic spectrum access scenario, where an unlicensed secondary user (SU) transmits a packet using the radio spectrum licensed to a primary user (PU). Considering a cognitive radio network, it is assumed that during the transmission period of an SU multiple interruptions from PUs may occur, increasing the time needed to transmit a packet. Assuming that the SU's packet length follows a geometric distribution, we start by deriving the probability of an SU transmitting its packet when k > 0 periods of PU's inactivity are observed. As the main contribution of this paper, we derive the characteristic function of the service time, which is further used to approximate its distribution in a real-time estimation process. The proposed methodology is independent of the SU's traffic condition, i.e., both saturated or non-saturated SU's traffic regime is assumed. Our analysis provides a lower bound for the service time of the SUs, which is useful to determine the maximum throughput achievable by the secondary network. Simulation results are used to validate the analysis, which confirm the accuracy of the proposed methodology.