Model-based Bandwidth Estimation in IEEE 802.11 Wireless Networks

Predicting the bandwidth achievable by a new flow is crucial for many bandwidth-aware services. In IEEE 802.11 based networks, the achievable bandwidth of a new flow is not necessarily the same as the free bandwidth on the channel due to the contention-based channel access method. In this paper, we develop a novel analytical model that provides closed-form calculation of bandwidth allocation in all possible operating states of single-hop IEEE 802.11 networks. Using this model, we design an algorithm that accurately predicts the achievable bandwidth of a new flow for single-hop wireless networks, such as wireless LANs, with or without service differentiation. In addition, we extend the bandwidth prediction model to multihop environments. Although the accuracy of the our bandwidth prediction model is degraded in multihop networks, our simulation results show that our algorithm still predicts more accurately than current existing methods