MAMoC-Android : Multisite Adaptive Computation Offloading for Android applications

Abstract

Computational offloading has been widely used to improve the performance of mobile applications and conserve the energy of mobile devices. Prior studies have primarily focused on a form of offloading where only a single server is considered as the offloading site. However, mobile devices now have access to a range of nearby mobile and fixed devices and multiple cloud providers. This paper proposes a method for multisite computation offloading in dynamic mobile cloud environments, in order to save energy and improve application execution time. Our proposed dynamic offloading decision algorithm takes into consideration the offloading score and records of past offloading executions to select the best candidate(s) for offloading. Multisite offloading execution achieves a greater reduction with respect to the completion time and energy consumption of mobiles when compared to local execution or a single-site offloading execution on a public cloud instance.