Enhancing mixed-grid optical switching networks: A dual-phase approach for resource optimization and security analysis

Abstract

Network operators typically find it challenging to upgrade their network infrastructure due to concerns about cost and service level agreements, particularly when it comes to backbone optical switching networks (OSNs). These days, network operators use backbone OSNs’ flex-grid to fixed-grid node migration process to support a multitude of bandwidth-demanding applications. But without careful planning, it could result in the wasteful use of resources. Also, during the migration process, the networks face security issues specific to optical communication networks including susceptibility to eavesdropping, data interception, unauthorized access, and denial-of-service attacks that compromise data confidentiality, integrity, and availability. This work offers resource allocation optimization methods for mixed-grid OSNs to optimize resource utilization. Modern optical networks feature complex architectures and a variety of technologies, making network management and information distribution challenging. This complexity is exacerbated by diverse optical technologies and service delivery protocols. This research addresses security and resource allocation in optical communication networks using a unique method that combines Siamese Heterogeneous Convolutional Neural Networks (SHCNN) with Triangulation Topology Aggregation Optimizer (TTAO). The introduced method consists of two phases. In the first phase, SHCNN-TTAO is proposed for Resource allocation. In the second phase, Software-Defined Fuzzy Alpine Skiing Neural Network for security analysis. Key performance metrics such as accuracy, Mean Square Error (MSE), Mean Absolute Percentage Error (MAPE), Root Mean Square Error (RMSE), and Mean Absolute Error (MAE) are comprehensively assessed. The proposed method attains a higher accuracy of 99.7%, and lower RMSE of 0.015329, MSE of 0.000235, and MAPE of 0.000343.