Dual optimization for enhancing TRA in MCF-based SDM-EONs

Abstract

A fine-grained, flexible frequency grid of elastic optical transmission and space division multiplexing (SDM) in conjunction with spectrally efficient modulations is an ideal solution for the impending capacity crunch. The routing, modulation, core, and spectrum assignment (RMCSA) problem is an important lightpath resource assignment problem in SDM elastic optical networks (SDM-EONs). Intercore Crosstalk (XT) degrades the quality of parallel transmissions on adjacent cores, and the RMCSA algorithm must satisfy XT requirements while optimizing network performance. There is an indirect trade-off between spectrum utilization and XT tolerance; while higher modulations are more spectrum-efficient, they are also less tolerant of XT because they allow fewer connections between neighboring cores on the overlapping spectrum. In the presence of XT, the Tridental Resource Assignment algorithm (TRA) has been shown to optimally assign resources in multicore fiber networks. Using the tridental coefficient (TC), this algorithm strikes a balance between spectrum utilization and XT. However, TRA is computationally expensive because TC is calculated for all the possible candidate resource sets for each lightpath request. In this paper, we demonstrate that acceptable or better TRA performance can be achieved while reducing computational overhead than traditional TRA. We achieve that using different stages of offline optimization, required only once. We demonstrate that by using tailored weights in TC, the performance of TRA can be enhanced further. We observe that acceptable TRA performance can be achieved with as low as 40% of the total computations and that adjusting weights reduces the bandwidth blocking of TRA, which is already performing better than the baseline algorithms, by approximately two orders of magnitude.