A Novel Model to Computational Offloading on Autonomic Managers: a Mobile Test Bed

Abstract

The technological advance of mobile devices, networks and cloud technologies progressed and reduced their access costs to the whole human population. Besides that, mobile devices are still limited in the battery capacity, storage, and connectivity. An efficient way to manage their resources is to make the applications self-adaptive and context-aware using the MAPE-K loop model. However, even this method can add a considerable processing cost to their devices. This paper proposes to reduce such costs by applying the computational offloading technique into the classical MAPE-K loop. In this way, we analyzed it based on literature evidence to find a suitable process that allows offloading to remote and cloud servers. The results through experimentation on the proposed model show that there is a substantial performance increasing in the planning activity remote executions compared to the local ones, what is also affected by the distance from the servers.