Accurate Power Monitoring at Operating System Level towards Energy-Aware Android Devices

Nowadays, mobile devices are more and more widespread thanks to their flexibility and advanced functionality. They are probably the device we use the most, since they permit us to access to many services (e.g. mail, Internet, social networks, etc.) when we are outside or we do not have another device, like a computer, close at hand. However, if on one hand mobile devices have many advantages, on the other hand, they have drawbacks as well. One of their main drawbacks is certainly power consumption, and this makes smart power management the key point in the enhancement of the user experience. Even though there exist many approaches to the problem, they are often inaccurate in estimating the remaining Time to Live (TTL) of the device. In order to estimate the TTL, it is necessary to gather information about the state of the device. Using an approach based on the device Operating System (OS), we can easily monitor the hardware components that mainly impact on the battery life. In the Android context, there are many applications (Apps) that provide information about the device TTL in this way. However, as far as they are part of the system, they may have a non-negligible impact on the battery life itself. Moreover, in particular cases, it could be necessary to collect such information with a certain frequency, in order to build reasonably efficient and accurate models. This work proposes an efficient power monitoring approach, deeply fused within the Android OS, able to provide accurate and reliable information about the device status and with a negligible impact on the battery. Most importantly, this work can be applied to contexts where a high frequency data gathering is required and essential for that particular purpose. This work is organized as follows: • in Chapter 1 we define the context of our work and briefly describe its motivations • in Chapter 2 we provide a general overview of Android OS, present its main features and characteristics and analyze its worldwide diffusion • Chapter 3 shows the most important works in the mobile devices power management field • Chapter 4 discusses the problem we want to address and our solution. MPower App is also presented • in Chapter 5, starting from Android architecture, we explain how our work is implemented, based on the findings in the previous chapter • Chapter 6 reports the results obtained testing our work from different points of view • in Chapter 7 we discuss the results together with the limits of our solution and with other researches possibly deriving from this work.