Article Preview
Top1. Introduction
As networks and connected devices become more mobile and thus energy constrained, and as the requirements for lower energy consumption become more demanding, the issue of consumed power by wireless network cards is becoming an intensively researched topic. Furthermore, concerns by the public on electromagnetic fields increase the interest of such research. When using a wireless connection, increasing the strength of a signal may result in better quality, as perceived by the end user application, but it also results to more energy consumption and stronger electromagnetic fields. On the other hand, if the signal strength is reduced too much, there is the risk of reduced performance or even losing connectivity completely.
Development of mechanisms that adjust the power consumption to optimal levels is considered necessary (Bornholt, Mytkowicz, & McKinley, 2012). For example, the authors in (Mayo & Ranganathan, 2003) conclude that, in order for single general-purpose mobile devices that combine multiple functionalities to achieve longer battery life, they should be designed to include requirements-aware energy scale-down techniques.
One efficient method to deal with this problem is to optimize the power consumption of network adapters. Research in devices such as smartphones has shown that WiFi adapters are responsible for a significant percentage of the consumed power, and thus energy savings in this area are very important for devices with limited power sources such as battery operated ones (Agarwal, Schurgers, & Gupta, 2005; Carroll & Heiser, 2010; Friedman, Kogan, & Krivolapov, 2011; Malik, 2012; Pering, Raghunathan, & Want, 2005; Rozner, Navda, Ramjee, & Rayanchu, 2010; Shye, Scholbrock, & Memik, 2009). For example, the author in (Malik, 2012) lists WiFi as the most power-consuming mode of an Android phone’s functions. Furthermore, research in (Carroll & Heiser, 2010) shows that under specific benchmarks the WiFi adapter can exceed 700mW in power consumption. Similar results have also been verified by measurements on smartphones (Balasubramanian, Balasubramanian, & Venkataramani, 2009) including measurements by end users taken “in the wild” (Shye, Scholbrock, & Memik, 2009).
The IEEE 802.11 standard deals with this problem by defining two modes; the active mode, where the network adapter is awake and can receive data at any time and power save mode. Where it cannot receive or send any data, so the energy consumption is reduced in that state. In (Zheng & Kravets, 2005), an on-demand power management technique is taking advantage of the above, to achieve 50% less energy consumption. In (Kravets & Krishnan, 2000), a transport layer mechanism enables the interface periodically or when necessary, reducing the energy consumption to 17%. However, the above mechanisms are based on active and inactive periods of the interface, which leads to additional delay at the arrival of the frames and degradation of the connection quality.
Another approach to the issue of optimizing power consumption has been made by focusing on other features of state-of-art standards, such as Multiple Input-Multiple Output (MIMO) operation. A mechanism focusing on radio chain management in order to reduce the power consumption has been proposed in (Yu, Zhong, & Sabharwal, 2009), which improved energy efficiency by 32% in best case scenario, with a high data rate (50Mbps). Another work related to MIMO utilization is in (Jang, Hao, Sheth, & Govindan, 2011), where the mechanism introduced, schedules the wake-up and sleep times of the clients, and on top of that, determines the antennas that are going to transmit in a way that is considered the most efficient.