Efficient Edge Nodes Reconfiguration and Selection for the Internet of Things

Abstract

Recent technological improvements have modernized the industrial sector. These improvements range from autonomous to automation of industrial processes. The industrial sector requires all the process to be accomplished locally because of the security and delay constraints. To accomplish the goal, edge/fog is a possible middleware between a cloud and an industrial environment. Edge nodes can provide processing with acceptable security and latency to robots, sensors, and actuators in the industrial environment. The industrial system configures various services at the edge nodes to enhance and automate the performance of the system. This paper considers an important but less investigated service hosting problem, where the edge nodes are dynamically reconfigured to host possibly the most recently requested services from the sensor nodes. Because of the limited storage and computational resource at the edge nodes, the problem of reconfiguration is important, which can increase the number and types of services hosted by the edge node. When a sensor node is within the transmission range of multiple fog nodes, to efficiently select the most appropriate fog node for data transmission, different types of fog node selection methods, random selection, shortest estimated latency first, and shortest estimated buffer first, have been considered and evaluated in this paper with satisfactory results.