Throughput enabled rate adaptation in wireless networks

Designing a rate adaptation algorithm that performs well in diverse settings is challenging due to the complex physical-layer effects of wireless links, including interference, attenuation, and multi-path fading. We present Throughput Enabled Rate Adaptation (TERA), a new approach to rate adaptation that accounts for interference and congestion effects implicitly but is based solely on measurements of the throughput attained at different data rates. TERA is surprisingly simple and compatible with existing 802.11 implementations. Extensive simulations and real-world experiments are used to show that TERA consistently outperforms all prior rate adaptation schemes used to date.