|Abstract:||Energy conservation is one of the greatest challenges in multi-hop wireless networks due to the ever-increasing energy requirements of wireless devices and the slow advancement of battery technology. While significant energy savings can be obtained by incorporating energy efficiency into the design of network protocols, the approaches taken so far have been very diverse. Current research has focused on either optimizing the energy use for a given communication task, or optimizing the energy consumption when the network is idle. However, an integrated approach is lacking. To this end, we explore the limits of traditional stand-alone techniques and expose some commonly held myths about energy conservation in wireless multi-hop networks. The main goal of our research is to develop a unified design that enables energy-efficient network operation.
In the first part of the thesis, we explore the existence of an optimal operating point that minimizes energy while satisfying the communication requirements on the network. Our goal is to conserve energy by accounting for all sources of energy consumption: (1) energy consumed for communication including energy spent for data and control overhead and (2) energy consumed during idling. Essentially, this is an energy-efficient network design problem. Since this problem is a node-weighted buy-at-bulk problem, which is NP-hard, we follow a divide-and-conquer approach, and first propose an on-demand topology management protocol, TITAN to reduce idling energy consumption. TITAN serves as a building block for a two-stage approach to energy-efficient network design, which first reduces the energy consumed in idling and second the energy consumed in data communication. Our results show that this two-stage approach is the only feasible approach that meets the challenge of operating the network with low energy cost without degrading communication.
From our experiences with energy-efficient network design, we observed that : (1) idling energy consumption should be the primary target for energy conservation, (2) any solution to energy conservation becomes cost prohibitive with increasing overhead, and (3) minimizing energy conservation does not necessarily improve network lifetime. Therefore, in the second part of this thesis, we build on our results to formulate solutions to each of these problems. To reduce idling energy consumption, we propose a bulk-communication protocol that achieves high energy savings from using a high-power high-rate radio for data communication and a low-power low-rate radio for network maintenance. To maintain energy-efficiency in the presence of high control and data overhead, respectively, we present two protocols: (1) adaptive recovery to maintain energy-efficiency in the presence of failures and (2) probability-based broadcast forwarding. Finally, to balance energy consumption in the network, we use a preemptive recovery protocol that re-distributes traffic based on the remaining energy of current forwarding nodes. Essentially, our results show that high energy savings, high communication performance and long lifetime in the network can be achieved through low-overhead, low-complexity protocols that rely on local decisions.