Adaptive Mini-Slot Selection for Low-Overhead ACP-2D Scheduling in Multi-OLT Multi-Lane PON-Based 5G Fronthaul

Abstract

Multi-optical-line-terminal (multi-OLT) multi-lane passive optical networks (PONs) are promising candidates for converged optical and wireless access, especially when fifth-generation (5G) fronthaul traffic must coexist with Internet of Things (IoT), fiber-to-the-home (FTTH), and wireless sensor network (WSN) services. The access class priority-based two-dimensional scheduling (ACP-2D) algorithm has shown that 5G fronthaul traffic can be transported over a 100 Gb/s next-generation Ethernet PON (NG-EPON) while satisfying the 250 μs 3GPP fronthaul delay requirement. However, a fixed mini-slot duration creates a direct tradeoff: shorter mini-slots reduce delay but increase overhead, whereas longer mini-slots reduce the overhead-to-data ratio (ODR) but may violate the latency bound. This paper proposes Adaptive ACP-2D, a lightweight extension that selects the largest feasible mini-slot duration before running ACP-2D. The proposed selector uses a delay constraint and the optical network unit (ONU) distance to choose between 125 μs and 71 μs mini-slots while preserving a 36 μs fallback mode. Numerical evaluation based on the published ACP-2D parameters shows that the proposed method maintains 5G delay within 250 μs over a 20 km reach and reduces the average 5G ODR by 56.83% compared with fixed 36 μs operation. The approach requires no physical-layer modification and adds only constant-size decision overhead to the existing ACP-2D scheduler.