Flow-based high throughput path selection in 802.11-based multi-rate wireless mesh networks

Today´s 802.11-based wireless network interface cards (WNICs) can leverage multiple transmission rates to exploit various and dynamic channel conditions. Some link adaptation schemes (LASs) in the literature have been shown to adapt well to dynamic channel conditions, thereby achieving high link throughput. However, routing protocols unaware of the underlying LAS may achieve low end-to-end throughput due to inappropriate path selection. In this paper, we propose a cross-layer routing framework called EAB which aims at discovering a high throughput path using statistics from MAC and LAS layer for a newly initiated flow. Our simulation results reveal that EAB operating on top of some efficient LASs significantly outperforms existing routing protocols in terms of the achieved throughput and the impact on existing flows.