Heuristic fragmentation-aware scheduling for a multicluster time-sensitive passive optical LAN

Abstract

With the development of Industry 4.0, optical networks are playing a key role in carrying large amounts of data traffic in industrial networks. Multicluster time-sensitive passive optical local area networks (POLs) use passive optical networks (PONs), which are important optical network segments, as local area networks (LANs); such networks can provide high-capacity real-time communications for large-scale industrial systems and have the potential to replace legacy LANs in industrial networks. To satisfy dynamic application requirements, POLs for future industrial networks need to be sufficiently flexible to adapt to frequent changes. Hence, this paper proposes a heuristic fragmentation-aware scheduling (FAS) method based on fast list scheduling for the scheduling of real-time traffic in multicluster time-sensitive POLs. The FAS algorithm selects a scheduling plan in accordance with a fragmentation calculation, schedules one time-triggered (TT) flow at a time from the pending scheduled TT flow set, and incrementally updates the scheduling table. The list processing mechanism of list scheduling is modified to improve scheduling performance, while the fast speed of list scheduling is inherited. Simulation results show that compared with a basic rate monotonic scheduling (RMS) list scheduler, FAS significantly increases the schedulable scale and schedulable load of the TT flow set while markedly reducing the TT communication overhead. With the help of FAS, multicluster time-sensitive POLs can support more critical tasks with higher network efficiency, allowing them to better cope with future dynamic and complex industrial application scenarios.