On multihop wireless network management: measurement, modeling and control

Multihop wireless networks are becoming a new attractive communication paradigm owing to their low cost and ease of deployment. Managing multihop wireless networks, however, is especially challenging due to a fluctuating wireless medium, presence of wireless interference, and the increasing demand for them to scale to large sizes. This dissertation tackles the multihop wireless network management challenges by systematically integrating measurement, modeling and control. On the measurement and modeling side, this dissertation develops a novel probabilistic region-based localization algorithm to accurately determine node locations with limited and noisy measurement information. The dissertation further develops a general model of wireless interference to estimate throughput and goodput between arbitrary pairs of nodes in the presence of interference from other nodes in a wireless network. Our model advances state of the art in interference modeling by (i) estimating interference among an arbitrary number of senders, (ii) modeling unicast transmissions, and (iii) modeling the general case of heterogeneous nodes with different traffic demands. On the control side, we investigated one of the most important network control problems—design of routing protocols for large wireless networks. We developed a new routing protocol, Small State and Small Stretch (S4) to jointly minimize routing state and routing stretch. S4 uses a combination of beacon distance-vector based global routing state and scoped distance-vector based local routing state to achieve a worst-case stretch of 3 using O( N ) routing state per node in an N-node network. Its performance benefits are further demonstrated in extensive simulation and testbed experiments.