Can execution time describe accurately the energy consumption of mobile apps? An experiment in Android

Measuring the energy spent by a software application is a problem that can be solved by having the proper hardware or software instruments. However, not always such tools are available or the provide resolution cannot fit the needs of the user, for instance when measuring a very small piece of code. This problem is particularly relevant on mobile software products, as they are developed to be executed in an environment limited in energy resources. Mobile s oftware engineers shou ld take special consideration on the energy consumption when designing and implementing an application. In this paper, we propose that the energy consumed by a unit of code can be approximated by the execution time. Using software benchmarks run with diffe rent data loads, we measured the execution time required to complete the job, and using a software tool to measure the energy spent during the execution of the benchmark, with the objective of finding a relationship among them. We observed that, regardless of the software benchmark, the data load injected and the programming language, of implementation the ratio between the execution time and the energy consumption remains consistent, opening the opportunity to develop techniques to approximate the energy c onsumption of mobile software based on execution time measurements