A non-preemptive mac protocol for multi-channel cognitive radio networks In this paper, we propose a novel multi-channel Medium Access Control (MAC) protocol for distributed Cognitive Radio Networks (dCRNs), which adopts a non-preemptive MaxWeight scheduling policy based on the availability of the licensed channels and the individual backlog of the Secondary Users (SUs). Under this protocol each SU competes for a data channel using the RTS-CTS mechanism, and the channel assignment is conducted by a distributed algorithm working without the need of a central entity. An analytical model for the individual SU’s transmission probability, packet service time, and network aggregate goodput is proposed, which relies on a Discrete Time Markov Chain used to model the SU MAC behavior. Simulation results considering different number of SUs and data channels are used to assess the proposed model, characterizing the level of idle spectrum utilization achieved by the MAC protocol.