Hybrid Deep Neural Network based Performance Estimation Method for Efficient Offloading on IoT-Cloud Environments

Abstract

The IoT-Cloud virtual machine system is a cloud-based execution solution for IoT devices with offloading techniques that delegate tasks requiring high computing power from low-performance IoT devices to a high-performance cloud environment as a service. The IoT devices with the IoT-Cloud virtual machine system can perform complex tasks using the computing power of high-performance cloud. The offloading technique can reduce the execution performance depending on the workload of the IoT devices and the clouds. Therefore, it is necessary to decide offloading execution considering the workload of the IoT devices and the clouds. In this paper, CPU utilization trend, which is one of the workload indices, is predicted through deep learning in order to decide offloading execution considering the workload of the IoT devices and clouds. In this paper, we present four CPU usage models and introduce a technique for predicting server load based on hybrid deep neural network. The predicted CPU utilization trend is indicative of future CPU utilization information and is therefore an indicator for offloading execution decisions. Through experiments, we confirmed that the proposed method estimates the load of the model very similar, and it can apply the offloading adaptively according to the load of the server.