Integrating SimEON with DeepRMSA for Dynamic Network Simulation of Elastic Optical Networks

Abstract In optical networks, simulations serve as a cost-effective and potent network planning and design method. Such simulations allow researchers and network architects to rapidly procure preliminary insights into network performance and facilitate easy design modifications. Regrettably, a significant portion of optical simulators remain proprietary. Presently, there's a notable scarcity of simulation tools in the optical network domain that harness the prowess of machine learning. Elastic Optical Networks (EON) offer a distinct advantage over their Wavelength Division Multiplexing (WDM) counterparts due to their refined channel spacing and versatile spectrum resource utilization, enhancing spectrum efficiency. Network resource allocation emerges as a paramount research focal point in this context. Within EON, this challenge is termed the Routing, Modulation, and Spectrum Allocation (RMSA) problem. Its objective is judiciously allocating network resources by pinpointing the optimal modulation format, ensuring call request fulfillment. SimEON stands out as a unique open-source simulation tool tailored for EON, adept at simulating an array of EON configurations and designing RMSA alongside regenerator placement/assignment algorithms. Moreover, it can be augmented with appropriate models to simulate CapEx, OpEx, and network energy consumption metrics. Deep learning (DL), a specialized branch of machine learning, leverages neural networks, extensive data sets, and algorithms to cultivate models adept at unraveling intricate challenges. In this paper, we extended the capabilities of SimEON by integrating the DeepRMSA algorithm into the existing simulator. We compared the performance of conventional RMSA and DeepRMSA algorithms and provided a convenient way for users to compare different algorithms' performance and integrate other machine learning algorithms.