Resumen
Executing tasks on mobile devices brings some challenges that
must be addressed, such as device heterogeneity, limited CPU power,
limited memory, short battery life, mobility, and intermittent disconnections.
BOINC (Berkeley Open Infrastructure for Network Computing), the
platform chosen in our research project that aims to process medical
images using mobile phones, does not completely solve these challenges. In
this sense, this article's objective is to present the modifications to BOINC's
scheduling strategy to consider the characteristics of mobile devices and
possible disconnections. The article describes BOINC-MGE (BOINC -
Mobile Grid Extension), a BOINC extension that incorporates two new task
scheduling algorithms to ensure the completion of tasks while efficiently
using the mobile devices' energy. The strategies are tested using an image
segmentation algorithm.
must be addressed, such as device heterogeneity, limited CPU power,
limited memory, short battery life, mobility, and intermittent disconnections.
BOINC (Berkeley Open Infrastructure for Network Computing), the
platform chosen in our research project that aims to process medical
images using mobile phones, does not completely solve these challenges. In
this sense, this article's objective is to present the modifications to BOINC's
scheduling strategy to consider the characteristics of mobile devices and
possible disconnections. The article describes BOINC-MGE (BOINC -
Mobile Grid Extension), a BOINC extension that incorporates two new task
scheduling algorithms to ensure the completion of tasks while efficiently
using the mobile devices' energy. The strategies are tested using an image
segmentation algorithm.
| Idioma original | Inglés |
|---|---|
| Publicación | Revista Eletr̂onica Argentina-Brasil de Tecnologias da Informação e da Comunicação |
| Volumen | 1 |
| N.º | 13 |
| DOI | |
| Estado | Publicada - 16 ene. 2021 |