Greedy-Based Edge Collaboration Scheme for Improving Quality of Experience

Abstract

Applications using the Internet of Things (IoT) are widely used in life and the number is increasing. These applications need lots of computation resources for service. However, IoT devices do not have enough computation resources. To solve this problem, edge server has been proposed. Edge server has the amount of computation resource than IoT devices and achieves low network delay. Nevertheless, the efficiency of task processing is degraded when the number of connected devices is large. For processing efficiency, collaboration schemes using other computing nodes have been studied. Collaboration schemes achieve high Quality of Experience (QoE) than local processing. However, the existing collaboration schemes do not consider computation or communication resources. Because these factors are not considered, existing collaboration schemes cannot select the optimal collaboration target for task processing. In this paper, we propose a greedy-based edge collaboration scheme for improving QoE. We first predict computation resource usage by the received task. Second, we determine collaboration based on the probabilistic model using the predicted resource. After the collaboration decision, finally, we collaborate based on the delay model according to the collaboration target. Experimental results show that the proposed scheme achieves high QoE compared to the existing collaboration schemes due to the success rate is high and completion time is low.