Optimizing offloading strategies for Mobile Edge Cloud systems

With the rapid growth of the number of mobile devices and the increase of the corresponding computation demand, it has been considered that Mobile Cloud computing and Edge computing will play the significant roles in the upcoming IoT era. It has become an active research topic to develop the offloading schemes for mobile devices, in which the tasks arriving at the mobile devices may be offloaded to run in the cloud or the edge devices. In this paper, mobile edge cloud systems are considered, which consists of mobile devices, edge devices and the cloud server, and the three-tier offloading schemes are proposed to achieve the optimal task performance in MEC. In the three-tier offloading schemes, the computation tasks arriving at the mobile devices may be offloaded to run on the edge devices while the edge devices may further offload the tasks to the cloud when the edge devices are overwhelmed. In this paper, two task modes are considered: batch mode and streaming mode. For the batch mode (i.e., the tasks arriving at the systems and being processed in batches), the offloading optimization problem is modelled as a Mixed 0-1 Integer Programming problem, aiming to minimizing the makespan of the batch of tasks. For streaming mode (i.e., the tasks arriving at the system continuously), the offloading optimization problem is formulated as a non-linear optimization problem, aiming to minimizing the average response time of a task in the task stream. The extensive experiments have been conducted to demonstrate the effectiveness of the proposed offloading schemes, and the impact of various parameters in the MEC systems is also evaluated