摘要：A Wireless Backhaul Network (WBN) carries aggregated data from multiple sources to core networks over wireless links. Increasing data rates and network densification have catapulted the IEEE 802.11 protocol as a strong candidate protocol for WBNs. The low-cost of deployment and maintenance, and self-management functionality are salient features of the IEEE 802.11 protocol for WBNs. Existing disc- graph models and signal-to-interference-and-noise (SINR) goodput models are widely used in sensor networks and mesh networks in simulation tools such as ns-2, ns-3, and Qualnet. However, these models have not been adequately studied in real world WBNs. Our empirical data show that in a typical WBN scenario, an irregular goodput pattern is neither conformant to the S1NR nor the disc-graph models. We find that topology and interferer proximity impact the WBN goodput at the MAC layer inconsistently leading to what we call MAC irregularity. We go on to identity the cause of MAC irregularity and quantify its impact on WBN goodput.
