Utilizing hybrid computing models for network monitoring and security analysis through optical network modeling and data analytics

Safe and reliable data transport over optical networks is essential for a high-speed Internet. Optical fibres are the backbone of the Internet, allowing billions of people around the world to connect and exchange data. However, these cables can be compromised by both deliberate attacks and accidental damage. The disruption caused by these abnormalities may result in significant monetary and data losses, undermine the security of optical networks by enabling unauthorised parties to view the sent data, or progressively lower network performance. Based on an examination of data from optical networks, this research proposes an original strategy for hybrid machine learning-based network monitoring and security improvement. Here, an anomaly detection method called convolutional principal component network is utilised to keep an eye on the optical infrastructure. Then, we apply hybrid cloud vector Bayesian graph networks to fortify the system. Anomaly detection rate, accuracy, precision rate, and quality of service are all experimentally evaluated across a variety of security analysis datasets. We validate our research using data from operational networks and simulated data created in a realistic-looking simulator. The findings verify the efficacy of our approach for optimising routing decisions in real time and designing networks for the future.