The effect of carrier sensing mechanisms on wireless mesh network goodput

For wireless mesh networks using the IEEE 802.11 family of standards, carrier sensing mechanisms such as physical carrier sensing and virtual carrier sensing play a key role in determining goodput and overall network performance. However, inaccurate modelling of the effect of carrier sensing on goodput leads to poor decision making in mesh network roll out. More accurate models can be derived by analysing network performance under non-ideal carrier sensing conditions which manifests as exposed nodes and hidden nodes. In this paper, we model the network goodput under both ideal and non-ideal carrier sensing by varying the transmitter-receiver separation and interferer proximity. Our results show that taking into consideration both transmitter-receiver separation and interferer proximity yields a four-phase goodput variation which have not been reported in the literature. The advantage gained by using distances rather than signal strength (or power) as the primary parameter is that our model will benefit the planning process of a wireless mesh network. The model based on distances can be directly integrated to practical planning tools such as MapInfo and MapBasic.