Energy efficient adaptation engines for android applications

Abstract

ContextThe energy consumption of mobile devices is increasing due to the improvement in their components (e.g., better processors, larger screens). Although the hardware consumes the energy, the software is responsible for managing hardware resources such as the camera software and its functionality, and therefore, affects the energy consumption. Energy consumption not only depends on the installed code, but also on the execution context (environment, devices status) and how the user interacts with the application.Objective In order to reduce the energy consumption based on user behavior, it is necessary to dynamically adapt the application. However, the adaptation mechanism also consumes a certain amount of energy in itself, which may lead to an important increase in the energy expenditure of the application in comparison with the benefits of the adaptation. Therefore, this footprint must be measured and compared with the benefit obtained.Method In this paper, we (1) determine the benefits, in terms of energy consumption, of dynamically adapting mobile applications, based on user behavior; and (2) advocate the most energy-efficient adaptation mechanism. We provide four different implementations of a proposed adaptation model and measure their energy consumption. ResultsThe proposed adaptation engines do not increase the energy consumption when compared to the benefits of the adaptation, which can reduce the energy consumption by up to 20%. ConclusionThe adaptation engines proposed in this paper can decrease the energy consumption of the mobile devices based on user behavior. The overhead introduced by the adaptation engines is negligible in comparison with the benefits obtained by the adaptation.

ProDSPL: Proactive self-adaptation based on Dynamic Software Product Lines

“…Additionally, we have shown with our experimental results that ProDSPL is able to provide more stable configurations of the system, requiring less reconfigurations and maintaining a good and regular quality of service. In these experiments, we do not consider the cost of reconfiguring the internal architecture of the system, but of course it introduces an important overhead in resource consumption (e.g., energy, cpu time), while it could interrupt user work during reconfiguration [31]. So, we consider that reconfiguration strategies that minimize the quantity of system adaptations required are very advantageous.…”

“…Minimizing the energy consumption of adapta- tion is specially important in battery powered devices, like mobile phones. In [31,32] it is shown that although the benefit of applying reconfiguration services is always worthy to maintain the good quality and health of the system, it introduces an energy consumption overhead that should be considered when choosing a self-adaptation strategy. Indeed, the authors of these papers compare three different mechanisms to execute a reconfiguration plan in Android phones, and the energy cost of monitoring and executing the plan goes from 0.68% to 2.30% ( [32], Figure 6) of the total cost (the decision making considered here is the most simplest one, rules based on if-clauses).…”

“…Any decrease in energy consumption of an app, would significantly mitigate the greenhouse effect of ICT technologies, when it is executed in thousand or millions of mobile phones. However, considering the energy consumption of software in mobile phones is very difficult to measure accurately [31], more experiments should be performed to achieve forceful conclusion, and this is beyond of the scope of this paper.…”

Supporting the evolution of applications deployed on edge-based infrastructures using multi-layer feature models

“…The task responsible for augmenting the reality are distributed among the user node, G1 and G2: the node of the user executes tasks t11, t13, t14, t17 and t19, using 195 Mb of RAM; G1 allocates t15, t16, t18, t20 and t21, reserving 210 Mb of RAM and 395 Mb of HDD; finally, G2 executes t12 and reserves 5 Mb of RAM and 20 Mb of HDD. Note that the user node has the entire application installed, delegating tasks according to the infrastructure characteristics and adapting its execution accordingly [10]. Thus, it does not reserve HDD.…”

A Tool for Energy Consumption Monitoring and Analysis of the Android Terminal

“…Current mobile terminals are equipped with a variety of I/O components, such as GPU, AI processor, AMOLED flexible screen, gravity sensor, ambient light sensor, and other sensors. Rich hardware brings people a richer and faster experience, while the problem is the increasing power consumption of terminals [8,9]. However, the battery technology of intelligent terminals has not made a breakthrough.…”