A simple model for non-homogeneous and non-saturated IEEE 802.11 DCF infrastructure BSS

We propose a simple unified analytical model to analyze the IEEE 802.11 DCF infrastructure BSS performance in terms of MAC delay, PLR, and throughput efficiency under non-homogeneous and non-saturation conditions. Our model accounts for: (i) asymmetric traffic load between an AP and its associated STAs of an infrastructure BSS; (ii) transition from the non-saturation to saturation mode (and vice-versa) from an AP perspective; (iii) heterogeneous traffic flows between STAs; and (iv) heterogeneous wireless channel conditions between BSSs of a multi-AP hotspot scenario, all in a single unifying framework. More specifically, we integrate a Markov chain model in conjunction with a finite queueing model to analyze the QoS performance of DCF infrastructure BSS, which will be useful for capacity analysis and the design of network control mechanisms. Extensive analyses and simulations have unveiled that the improper modeling or ignorant of backoff freezing for an infrastructure BSS will result in overly conservative bounds which will lead to low network utilization when deployed as admission control, particularly, in heavy load scenarios.