Routing, Protection, and Wavelength Assignment in Large-Scale WDM Optical Networks: Performance Trade-offs, Failure Resilience, and Emerging Hybrid Approaches

Authors:

G Praveen Babu¹, K V Ramana²

¹Research Scholar, Research Scholar, JNTUK, Kakinada, India

²Professor, JNTUK, Kakinada, India

³Third Author Department & College

Abstract - The routing of wavelength division multiplexed optical networks still constitutes the basis of high-capacity communication infrastructures in the modern world, but the complexity of their operation has grown significantly as networks in size, heterogeneity in traffic, and resiliency demands. Routing, protection, and wavelength assignment strategies, in this case, are critical in shaping the throughput, blocking likelihood, and recovery performance during steady and failure induced environments. This literature review presents a designative and to a certain extent, critical synthesis of three inseparable research directions in large scale WDM optical networks: performance- based routing and wavelength assignment, protection and survivability measures, and new paradigms of hybrid models that introduce classical algorithmic processes together with machine learning-based decision support. The research papers under survey are arranged initially based on the static and dynamic traffic models, and the impact of assumptions of demand predictability on the computation routes and wavelength utilization are emphasized. The protection strategies described as dedicated, shared, and pre-planned backup schemes are then reviewed keeping in mind recovery time, signaling overhead, and spectrum efficiency trade-offs witnessed in real world deployments. Special focus is given to hybrid methods combining offline optimization with learning-based backup path prediction, in which systems that have learned models based on past failures, make the effort, though not necessarily conclusive, to minimize connection disruption without necessarily over-reserving resources. In the analyzed sources, similar issues arise, such as scalability in the context of dense wavelength grids, the non-stationarity of results under non-stationary traffic, and a lack of reproducibility of the results related to using the ML-centric framework owing to the data limitations. The paper summarizes these findings to establish open research directions and design concerns that, based on the argument, are underutilized to date in the research on WDM network and provide a basis on which researchers and practitioners can explore the emerging world of resilient and wavelength-efficient optical networking. Meanwhile, the review has deliberately not presented any novelties in terms of algorithms, but rather given a comparative insight, methodological shortcomings, and practices of evaluation which impact on the realistic evaluation of future routing and protection structures in large-scale experimentation and operational WDM settings in contexts of worldwide investigations carried out in the present day.

Key Words:  WDM optical networks, routing and wavelength assignment, protection and survivability, failure-resilient optical routing, hybrid machine learning approaches