A Hybrid Task Crash Recovery Solution for Edge Computing in IoT-Based Manufacturing

Abstract

Smart factory is the deep integration of technology, business, products, and industry. The R&D of products, the construction of factory information, customer orders, and supply chain data are combined to act on the production. This paper proposes a Hybrid Internet of Things (H-IoT) platform framework, which consists of five layers: intelligent device layer, communication protocol layer, edge computing layer, IoT control layer, and application layer. Object Linking and Embedding (OLE) for Process Control (OPC) transmits the received device data to the database of the edge server through the TCP/IP protocol. After the edge server performs preprocessing such as data cleaning, the device data are uniformly collected and distributed through the IoT layer. The various subsystems of the smart factory obtain the required resource data through IoT. To run tasks uninterruptedly and efficiently, we propose a Container-Managed Task Crash Recovery Mechanism for the Edge Computing (EC) layer. We design a reasonable EC hierarchy, then propose a task migration scheduling strategy. Our method ensures the reliability and stability of manufacturing with mass data transmission even in the case of a random crash of partial edge servers.