“Exotic” routing protocol design and implementation for wireless mesh networks

Wireless mesh networks (WMNs) are increasingly being deployed for providing cheap Internet access. A main challenge in city-wide WMNs is to deal with the poor link quality due to urban structures and interference. To combat the high loss rates and satisfy the ever-growing demands for high-throughput Internet access, researchers have recently proposed a new class of “exotic” routing protocols exploiting techniques such as opportunistic routing or network coding, that promise a dramatic increase in overall network throughput. Nonetheless, the design of practical protocols exploiting such techniques still faces several challenges that often prevent them from realizing the maximum possible gains. This thesis presents the design, prototyping, and evaluation of two novel “exotic” routing protocols that address fundamental limitations of state-of-the-art protocols for multicast and unicast routing, respectively. The first part of the thesis presents Pacifier, the first reliable multicast protocol for WMNs that efficiently addresses the well-known (from the Internet) “crying baby” problem, i.e., when the presence of a poorly connected multicast receiver results in performance degradation for the rest of the receivers. The second part of the thesis presents a novel solution to the main challenge in network coding based opportunistic routing protocols, i.e., determining how many coded packets each forwarding node should transmit, through the design of CCACK. Finally, in the third part of the thesis, we review the current diverse practices in the evaluation methodology of this new class of “exotic” routing protocols, examine their strengths and their weaknesses, and make recommendations for more consistent and meaningful evaluation methodologies