• Sanjana
• Dr. Nitin Kumar

Building self-driving Software-Defined Networks requires network modeling, particularly to identify the best routing protocols that satisfy the objectives stated by administrators. However, the requirements for providing precise estimates of pertinent performance indicators like latency and jitter are not addressed by present modeling methodologies. In this study, we provide a unique Graph Neural Network (GNN) model capable of comprehending the intricate link between topology, routing, and input traffic to generate precise estimates of the mean delay and jitter for each source/destination pair. Because GNN is designed to learn and model data that is organized as graphs, our model can generalize over any topologies, routing protocols, and varying traffic intensity. Additionally, we demonstrate the model's potential for network operation through the presentation of a number of use cases that demonstrate its successful application in the optimisation of delay and jitter for each source-destination pair as well as its generalization abilities by reasoning in topologies and routing schemes that were not encountered during training.

Software Designed Network,Graph Neural Network,Neural Network

"Modeling SDN using Graph Neural Networks ", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.10, Issue 8, page no.f422-f427, August-2023, Available :http://www.jetir.org/papers/JETIR2308550.pdf

"Modeling SDN using Graph Neural Networks ", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.10, Issue 8, page no. ppf422-f427, August-2023, Available at : http://www.jetir.org/papers/JETIR2308550.pdf