This paper introduces the application of attentiontemporal graphic neural networks to enhance traffic flow and reduce delay in software-defined networks (SDN). Graphical Neural Networks (GNN), which have recently gained popularity for their efficiency in traffic data analysis, are further refined in attention temporal GNN by incorporating time as a critical variable. This paper emphasizes the role of each node within the network, representing individual data points, and the links that illustrate the interconnection and traffic intensity between them. The attention temporal GNN framework is constructed with multiple layers, each layer representing a unique time segment within the neural network. Central to this architecture are two critical variables in each layer: one indicating the state of the layer at a given moment and the other reflecting the traffic load at that specific point in time. By leveraging datasets generated from SDN environments, the GNN model is trained to enhance the network traffic management and optimization. This study demonstrates the effectiveness of the attention temporal GNN model in elevating SDN performance, marking a significant advancement in network management technology. Index Terms—Graphical Neural Network (GNN), SoftwareDefined Networking (SDN), OpenFlow, RouteNet.


    Access

    Check access

    Check availability in my library

    Order at Subito €


    Export, share and cite



    Title :

    Graphical Neural Network-Enabled Software-Defined Networking Technique for Naval SCADA Systems


    Contributors:
    Tomar, Shaivi (author) / Smith, Andrew (author) / Li, Yan (author) / Du, Liang (author)


    Publication date :

    2024-06-19


    Size :

    3785789 byte





    Type of media :

    Conference paper


    Type of material :

    Electronic Resource


    Language :

    English