Risk-Aware Cloud-Edge Computing Framework for Delay-Sensitive Industrial IoTs
Yi Zhang, Hung‐Yu Wei
Abstract
The industrial Internet of Things (IIoT) has been widely deployed to provide autonomous inspection on current production status and quality of products for modern manufacturing. However, the IIoT sensors generally are short of computing capabilities and therefore could not offer acceptable latency for computation-intensive inspection tasks. Besides, the mission-critical industrial applications are extremely sensitive to inspection failure, which may lead to serious manufacturing problems or accidents. In this paper, we propose a risk-aware cloud-edge computing framework for the delay-sensitive inspections of autonomous manufacturing. Due to the uncertainty of 802.11ax, we utilize the conditional value-at-risk (CVaR) to measure the inspection risk basing on the distribution of channel access delay. We develop a branch-and-check (BNC) approach to optimally and efficiently deploy the decomposable inspection tasks with the minimum operation cost and acceptable latency. The extensive simulations guide the operational use for future IIoT and the results show that the proposed system can save a large amount of unnecessary operation cost by enabling the processor sharing strategy.