BLAST: Battery Lifetime-constrained Adaptation with Selected Target in Mobile Devices

Abstract

Mobile devices today contain many power hungry subsystems and execute different applications. Standard power management is not aware of the desired battery lifetime and has no visibility into which applications are executing. However, power consumption is strongly dependent on which applications are executed. In this work, we propose a novel power characterization strategy for mobile devices called application-dependent power states (AP-states). Based on that, we formulate a management problem to improve performance under battery lifetime constraints, and we implement the management framework on a real Android device. We call our framework BLAST: Battery Lifetime-constrained Adaptation with Selected Target. The goal of such framework is to maximize performance while letting the device battery to last at least for a certain required lifetime, and only requires the user to select the desired target lifetime. The implementation does not require OS modifications and can be ported and installed to any Android device. We experimentally verify that our strategy can still meets user experience requirements with a selected target battery lifetime extension of at least 25%.

“Tour de France” data for the improvement of energy consumption in devices powered by limited energy sources

“…This dataset was used to model the energy consumption of one or more specific applications in a mobile device [6] , [7] . On the one hand, it allows the monitoring of energy consumption according to the frequency and the total number of operations, on the other hand, to put into practice our proposed methodologies through several concrete case studies.…”

Measurement-based methodology for modelling the energy consumption of mobile devices

“…Cinder's mechanism (Roy et al, 2011) consists of detecting the main energy-consuming tasks and then assigning this information to the guiding applications in order to allocate the energy resources to the highest priority tasks (Zeng et al, 2002(Zeng et al, , 2003Fonseca et al, 2008;Flinn and Satyanarayanan, 1999). Mercati et al (2015) have shown in battery lifetime-constrained adaptation with selected target (BLAST) that performance can be maximised if the user can select a target battery lifetime. This framework is an application-aware power management framework for mobile devices which controls operating conditions in order to meet a predefined battery lifetime.…”

Control algorithms for energy optimization in multimedia hand-held devices

“…Energy awareness is built into each application and depending on the battery state, which is monitored by a sensor. Mercati et al [25] proposes the Applications-dependent Power states (AP-states) to monitor the frequency of Central Processing Unit (CPU) and Graphics Processor Unit (GPU) as well as the execution time of each APP. Then, the average power consumption of each APP can be calculated by the pairs of frequency and execution time.…”

“…Kamat [11] built the energy awareness into each application and the battery lifetime is extended by changing the operating point of the applications. Mercati et al [25] presents a method to maximize performance of applications while letting the device battery to last at least for a certain required lifetime. In their research, the power consumption of frequently used APPs are measured to estimate the battery lifetime while running different APPs.…”