Feedback-Based Adaptation for Improved Power Consumption

In this paper we present a feedback-based adaptation mechanism that adjusts the transmission power of a wireless card on commodity PCs depending on the quality of the connection. Our purpose is to manage the available power in order to achieve lower power consumption without negatively affecting the user's perception of connection quality. We based our implementation on an existing theoretical model and focused on resolving problems and removing assumptions which made it inefficient in real life implementation. The initial model manages to minimize the power consumption in networks with exactly two nodes. In this paper, we extend the model to consider the possibility of the existence of a base station, where any number of nodes can be connected. Our objectives for the base station are to minimize the power consumption and guarantee continuous connectivity for all mobile nodes. We implement the adaptation mechanism for a specific adapter with open sources drivers thus allowing necessary modifications. We conduct a number of real world experiments. The results indicate that power consumption can be significantly reduced for nodes that are either almost stationary or slowly moving (e.g. at walking speed), without any significant increase in packet loss. The results are quite important as nowadays mobile devices with limited battery life time use tethering to become base stations for other devices like in ad-hoc networks.