Toward scalable passive optically powered fronthaul networks

Abstract

The evolution of wireless networks toward 5G and beyond has introduced challenges in designing efficient and cost-effective fronthaul networks. This paper focuses on the power-enabled optical fronthaul design (POFD) problem, which seeks to optimize the placement of optical splitters and their connections to radio units and distributed units while incorporating power over fiber (PWoF) technology. PWoF technology offers significant advantages by enabling simultaneous data and power delivery over optical fibers, thus eliminating the need for separate power units and improving network resilience. However, the POFD problem is NP-hard, as the complexity of determining optimal splitter placements and interconnections grows exponentially with the network size. To address this complexity, we propose two comprehensive heuristic-based approaches and benchmark their performance against clustering-based methods such as k-means clustering and genetic algorithms. For small networks, where exact solutions can be computed using integer linear programming (ILP), we demonstrate that our heuristics achieve near-optimal results. For large networks, where ILP is computationally infeasible, our heuristics provide scalable and effective solutions and maintain high performance compared to existing clustering-based methods in the literature that do not guarantee feasible solutions.