Improving the Performance of SDM-EON Through Demand Prioritization: A Comprehensive Analysis

• Main Authors: Patricia Morales, Astrid Lozada, Danilo Borquez-Paredes, Ricardo Olivares, Gabriel Saavedra, Ariel Leiva, Alejandra Beghelli, Nicolas Jara
• Format: Article
• Language: English
• Published: IEEE 2021-01-01
• Series: IEEE Access
• Subjects: Elastic optical networks; space-division multiplexing; resource assignment; network capacity; physical-layer impairments
• Online Access: https://ieeexplore.ieee.org/document/9410611/

This paper studies the impact of demand-prioritization on Space-Division Multiplexing Elastic Optical Networks (SDM-EON). For this purpose, we solve the static Routing, Modulation Level, Spatial Mode, and Spectrum Assignment (RMLSSA) problem using 34 different explainable demand-prioritization strategies. Although previous works have applied heuristics or meta-heuristics to perform demand-prioritization, they have not focused on identifying the best prioritization strategies, their inner operation, and the implications behind their good performance by thorough profiling and impact analysis. We focus on a comprehensive analysis identifying the best explainable strategies to sort network demands in SDM-EON, considering the physical-layer impairments found in optical communications. Also, we show that simply using the common shortest path routing might lead to higher resource requirements. Extensive simulation results show that up to 8.33% capacity savings can be achieved on average by balanced routing, up to a 16.69% capacity savings can be achieved using the best performing demand-prioritization strategy compared to the worst-performing ones, the most used demand-prioritization strategy in the literature (serving demands with higher bandwidth requirements first) is not the best-performing one but the one sorting based on the path lengths, and using double-criteria strategies to break ties is key for a good performance. These results are relevant showing that a good combination of routing and demand-prioritization heuristics impact significantly on network performance. Additionally, they increase the understanding about the inner workings of good heuristics, a valuable knowledge when network settings forbid using more computationally complex approaches.