Load Dependent Dynamic Path Selection in Multi-Radio Hybrid Wireless Mesh Networks

Abstract

Hybrid Wireless Mesh Networks provide improved connectivity and reliability over traditional wireless networks. However due to rapidly changing traffic patterns, such networks are more likely to experience congestion which leads to data loss. Inability to differentiate between congested links and select best possible paths dynamically can be very expensive in a rapidly changing network topology. This thesis proposes D-WCETT (Dynamic Weighted Cumulative Expected Transmission Time), an enhanced version of the WCETT routing metric, with focus on forming routes based upon network load information. The proposed routing metric dynamically selects paths with least level of congestion by taking into consideration locally available queue information. We present an evaluation of our implementation via extensive simulations and results indicate that the proposed metric significantly outperforms its predecessor by virtue of its ability to distinguish and dynamically select least congested paths in a multi-radio environment.