Online Offloading of Delay-Sensitive Tasks in Fog Computing

Fog computing, which provides low-latency computing services at the network edge, is an enabler for the emerging Internet of Things (IoT) systems. Task offloading is one of the main technologies of fog computing. The IoT devices with insufficient computing power will offload tasks to other devices with surplus resources to process. Those devices include IoT devices and fog devices. It exists a problem with how to find suitable devices to offload effectively. For this problem, this paper proposes a Short-Sighted-UCB (SS_UCB) algorithm based on the Upper Confidence Bound (UCB1) algorithm to perform one-to-many predictive offload, predicting which device can be offloaded to reduce task latency and improve quality of service (QoS). Furthermore, this paper proposes an Online-Learning-GS algorithm to solve many-to-many offload to minimize overall task latency. The experiments show that the effectiveness of the SS_UCB and Online-Learning-GS algorithm in a dynamic environment.