Modulation-Aware Asymmetrical Traffic Splittingfor Heterogeneous Demands

The efficiency of survivable elastic optical network (SEON) can be enhanced by optimizing traffic demand allocation and reducing the probability of blocking, especially in scenarios where heterogeneous traffic demands require dynamic resource allocation. One of the existing methods, stateaware asymmetrical traffic demand splitting (ASTD), splits the total traffic demand across link-disjoint multiple (LDM) paths with state-aware allocation strategies (SAAS). To further improve performance, this paper proposes a modulation-aware asymmetrical traffic demand (MASTD) splitting model that integrates distance-adaptive modulation (DAM) with SAAS to optimize the connection acceptance rates and reduce blocking probability (BP) for heterogeneous traffic demands. The MASTD-Mixed-integer linear programming (MILP) formulation is proven to be NP-hard, and to address scalability concerns, we propose a heuristic based First-Fit Demand Decreasing solution for efficient traffic demand allocation by maintaining survivability in EON. Simulation results on realworld networks show that the proposed model significantly enhances network performance, reducing BP and increasing spectrum utilization compared to the state-aware ASTD model.