Energy efficient schedulers in wireless networks: Design and optimization

Jeongjoon Lee, Purdue University

Abstract

Minimizing energy consumption is crucial for portable wireless stations because they operate on a limited battery supply. For example, the IEEE 802.11 standard includes a mechanism called power-saving mode (PSM), which allows a network interface on a mobile station to enter a sleep state whenever possible to reduce its energy consumption. We consider a generic wireless system composed of an access point (AP) and several stations that offer a PSM to its users. Our PSM is AP-centric (i.e., gives control to the AP) to save more energy. We formulate a downlink scheduling optimization problem aimed at saving energy and propose two heuristic scheduling policies to solve it. One of these policies is non-work-conserving, offering an interesting tradeoff between energy consumption and user performance. We next study and show how the length of the Beacon Period (BP) has a significant impact on the energy and the delay performance of wireless stations. For each of our two scheduling policies, we derive simple approximate formulas for the length of the BP that minimizes the energy consumption and for the relationship between the delay performance and the length of the BP. Assuming the maximum allowable average packet delay is given by the users as a QoS requirement, we illustrate how to dimension the length of the BP for the two schedulers we have proposed and show that we can achieve significant energy savings while meeting the delay constraint with the non-work conserving scheduler in many cases. We extend our ideas to the case when the channel condition is time-varying. We propose an opportunistic Power-Saving Mode and a corresponding scheduler design in wireless local area networks, which improves energy saving of the PSM stations while maintaining throughput maximization that is achieved by exploiting time-varying channel condition. We identify the challenges in the design and implementation of the PSM and the scheduler. We design and propose a channel probing scheme and a scheduler, which considers throughput maximization first and then focus on minimizing energy consumption. Extensive simulations show that our scheme saves significant amount of energy while maintaining throughput maximization.

Degree

Ph.D.

Advisors

Shroff, Purdue University.

Subject Area

Electrical engineering

Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server
.

Share

COinS