A dynamic reachable distance estimation method for spectrum assignment problem in MB-EONs considering nonlinear interferences

Multi-band elastic optical networks (MB-EONs) are a promising technology for future backbone networks. Routing, band, modulation level and spectrum assignment (RBMSA) problem is one of the key issues in MB-EONs. Due to the significant increase of total power in fiber links, nonlinear interference (NLI) must be considered when determining the highest available modulation level. Generally, two NLI-aware modulation selection methods are used in existing works, namely the AccurateSNR and FixedReach methods. The former is with high computational cost, while the latter is with low spectral efficiency. In this paper, we propose a novel NLI-aware modulation-selection method named dynamic reachable distance estimation (DRDE). The DRDE method can dynamically estimate reachable distances of modulation levels under realtime NLI by a simple formula. We further design a DRDE-based RBMSA algorithm to demonstrate the usage of the DRDE method for modulation level selection. Simulations show that the DRDE-based RBMSA algorithm outperforms the FixedReach-based RBMSA algorithm in terms of bandwidth blocking probability, and outperforms the AccurateSNR-based RBMSA algorithm in terms of average running time. Hence, the proposed DRDE method is not only computationally efficient but also spectrally efficient.