Closing the gap between traffic workload and channel occupancy models for 802.11 networks

The modeling of wireless network traffic is necessary to evaluate the possible gains of spectrum sharing and to support the design of new cognitive protocols that can use spectrum efficiently in network environments where diverse technologies coexist. In this paper we focus on IEEE 802.11 wireless local area networks and close the gap between two popular levels of modeling, macroscopic traffic workload modeling and microscopic channel occupancy modeling. We consider traffic streams generated by established traffic workload models and characterize the networking scenarios where a simple, semi-Markovian channel occupancy model accurately predicts the wireless channel usage. Our results demonstrate that the proposed channel occupancy model can capture the channel idle time distribution in most of the scenarios, while the Markovian assumption cannot be validated in all cases.