Joint Orchestration of Content-Based Message Management and Traffic Flow Steering in Industrial Backbones

Abstract

The industrial internet of things has radically modified industrial environments, not only enabling novel services but also dramatically increasing the amount of generated traffic. Nowadays, a major concern within industrial plants is to support network-intensive services, such as real-time remote vibration monitoring of autonomous guided vehicles, while ensuring the prompt and reliable delivery of mission-critical safety-related messages among machines and the control room. To this purpose, we present a novel solution jointly orchestrating content-based message management and traffic flow steering: the former enables edge-powered in-network processing modules to process packet payloads as they traverse the industrial backbone, the latter supports dynamic (re)routing of traffic flows towards such processing modules. In particular, we exploit software-defined networking for flexible traffic flow (re)routing and Kubernetes for dynamic deployment on edge nodes of in-network processing modules for content-based message management. As demonstrated by performance results based on our working proof-of-concept prototype, our solution efficiently allows to manage industrial traffic flows in a coordinated fashion, by considering requirements of concurrently running industrial applications and the current state of the overall topology.