Investigating the Impact of Fragmentation on Blocking Performance in Spectrally-Spatially Elastic Optical Networks

Spectrally-spatially elastic optical networks (SS-EONs) face challenges such as fragmentation and crosstalk, which lead to a rise in blocking probability. Accurate modeling of fragmentation and blocking probability is complex due to multiple constraints in SS-EONs. We evaluate the average fragmentation without considering spectrum contiguity constraints while simultaneously avoiding inter-core and inter-mode crosstalks in SS-EONs, which has not been studied in previous research. This study employs a continuous-time Markov chain-based analytical model to compute average fragmentation and blocking probability under different spectrum allocation policies—core-mode-spectrum random fit (CMS-RF) and core-mode-spectrum first fit (CMS-FF)—without contiguity constraints while simultaneously avoiding inter-core and inter-mode crosstalks in SS-EONs. The model considers both single-class and multi-class requests, where single-class requests use a uniform number of slots, and multi-class requests have variable slot requirements. Additionally, an iterative approximation model is introduced to improve scalability in single-hop links. Numerical results show that both the analytical model and simulations demonstrate lower blocking probability when contiguity constraints are not enforced, with CMS-FF performing better than CMS-RF in reducing fragmentation.