Data-driven dynamic load balancing with deflection routing for high-capacity wavelength division multiplexed optical network

Wavelength division multiplexing (WDM)-based optical networks (ONs) are essential for supporting the growing demand for high-capacity data transmission in modern communication systems, where efficient traffic management and load balancing are critical to ensure optimal resource utilization and minimize congestion. Traditional routing approaches, such as static and heuristic methods, often fail to adapt to dynamic traffic conditions, leading to inefficient bandwidth usage and degraded performance. This paper proposes an intelligent traffic-aware load balancing framework for WDM optical networks that models the network as a graph and dynamically distributes traffic based on real-time network conditions. By analyzing traffic patterns across nodes and WDM-enabled links, the framework selects optimal routing paths to balance load across wavelengths. Simulation results demonstrate that the proposed method achieves significant performance improvements, including an increase in throughput up to 90 Gbps compared to 61.1 Gbps in static routing, a reduction in packet loss rate to 0.017 from 0.051, and improved load balancing efficiency reaching 0.94. Additionally, network latency is reduced to 343 ms, indicating enhanced responsiveness. These results confirm that the proposed approach effectively mitigates congestion, improves link utilization, and enhances overall network stability under dynamic traffic conditions.