EstiTO: An Efficient Task Offloading Approach Based on Node Capability Estimation in a Cloudlet

In recent years, computation offloading over a cloudlet, which consists of a set of near-user devices, is recently studied to expand the limited resources on mobile devices. The challenge is that the compute nodes in a cloudlet typically have heterogeneous capabilities. Thus, the job execution time can be prolonged if the tasks are not appropriately assigned to these nodes. In this paper, an efficient task offloading approach is proposed. The proposed approach can assign tasks to nodes according to the node capability estimation. Because the objective is to achieve load balancing, the job execution time can be reduced. The experimental results show that the job execution time can be efficiently reduced by the proposed approach.