Towards Optimising WLANs Power Saving: Novel Context-aware Network Traffic Classification Based on a Machine Learning Approach

Abstract

Energy is a vital resource in wireless computing systems. Despite the increasing popularity of Wireless Local Area Networks (WLANs), one of the most important outstanding issues remains the power consumption caused by Wireless Network Interface Controller (WNIC). To save this energy and reduce the overall power consumption of wireless devices, most approaches proposed to-date are focused on static and adaptive power saving modes. Existing literature has highlighted several issues and limitations in regards to their power consumption and performance degradation, warranting the need for further enhancements. In this paper, we propose a novel context-aware network traffic classification approach based on Machine Learning (ML) classifiers for optimizing WLAN power saving. The levels of traffic interaction in the background are contextually exploited for application of ML classifiers. Finally, the classified output traffic is used to optimize our proposed, Context-aware Listen Interval (CALI) power saving modes. A real-world dataset is recorded, based on nine smartphone applications’ network traffic, reflecting different types of network behaviour and interaction. This is used to evaluate the performance of eight ML classifiers in this initial study. Comparative results show that more than 99% of accuracy can be achieved. Our study indicates that ML classifiers are suited for classifying smartphone applications’ network traffic based on levels of interaction in the background.