CO-scheduling and routing in a hybrid TSN and TDM-PON network for industrial flows

Abstract

With the development of smart manufacturing, information technology (IT) and operation technology (OT) are gradually converging to carry various types of industrial flows that are characterized by time sensitivity and bandwidth hunger. To cope with their communication needs, hybrid networking technology has become a trend, where integrating time-division multiplexed passive optical networks (TDM-PONs) with time-sensitive networks (TSNs) is garnering considerable attention. TDM-PON offers enhanced bandwidth for north–south communications, while TSN delivers greater flexibility for east–west communications. However, this integration presents challenges related to delay in end-to-end (E2E) industrial flows traversing both the TSN and PON domains. Traditional scheduling approaches only guarantee deterministic transmission within each domain that leads to reduced schedulability and inefficient resource utilization. To address these issues, we propose a cooperative scheduling (CO-scheduling) and routing scheme that spans both the TSN and PON domains, aiming for a global optimization of E2E industrial flows. Initially, we introduce an Optical TSN model, which serves as a TSN equivalent for TDM-PON. By treating TDM-PON and traditional TSN switches as a unified network domain, we can implement CO-scheduling and routing. This CO-scheduling approach can facilitate E2E deterministic transmission by considering the delay budget as a whole. We utilize time-aware transmission window and cyclic transmission window (C-TW) strategies that were proposed in our previous works to ensure the determinism of synchronous and asynchronous periodic flows, respectively. Moreover, we propose an integer linear programming (ILP) model and a “zero-waiting” heuristic algorithm to enhance schedulability and resource utilization efficiency. Simulation results show that the average schedulability and resource utilization efficiency of the CO-scheduling scheme are up to 20.83% and 41.37% higher than that of the traditional non-cooperative scheduling scheme, respectively.