Adaptation to hardware variability for energy efficient use in the internet of things

Resumo

Due to the evolution of silicon chips manufacturing processes at smaller and smaller nanoscales, the computing and memory substrates are increasingly changing, especially from an energy point of view. Considering projects with transistors composed of countable atoms quantities, small variations in the construction process can lead to energy costs of 10 times greater for the same processing. Software systems that identify hardware characteristics and take measures to adapt to the execution context amortize unnecessary energy expenditures. This economy is crucial for the execution continuity of embedded systems and/or applications used in the Internet of Things (IoT), where the energy resources are limited. One way to do this is by staggering tasks efficiently, postponing low-urgency services and running them all in only one period of CPU activity. The objective of this project was to increase the energy efficiency of IoT devices by creating a system that is capable of reorganizing and scheduling its tasks, as well as redistributing them in a network to be executed in nodes with characteristics of lower resource consumption.