Design And Implementation Of Intent-driven SDN Traffic Awareness System

The idea of separating the control plane and the data plane of the SDN network can realize the management and control of the network.With the deepening of SDN network research,how to simplify user operation and use of the network and improve the degree of network automation has become a new problem.Intent can quickly capture the user's thoughts,realize network networking and configuration with the help of SDN,and realize a simpler operation mode,which has become a new hot topic in current research.Network traffic awareness can obtain the current network status in time,accurately perceive network changes,and maintain stable network operation.Therefore,the thesis combines intent with network perception,and the main work is as follows:First,the overall framework of the system is designed.The system consists of an intent module,a data acquisition module,an SDN control module,a traffic prediction module and an interface.Each module cooperates with each other.The SDN control module captures the information from the intent module and responds to the data plane.The data acquisition module is used to acquire and store data.The traffic prediction module trains the neural network model and realizes real-time prediction of data.With the help of NLP,the intent translation algorithm is designed,the measurement strategy knowledge base is established to complete the intent mapping,the connection with the SDN network is established through the intent northbound interface NBI,and the intent interaction is completed.After the network traffic indicators are defined,the intentions related to "measurement" are issued,and the collection and storage of network,link,node and other related data are realized step by step,and the measurement results are displayed on the system visual interface.In view of the burstiness and spatial correlation of current network traffic time,the thesis extracts spatial features through graph convolutional network GCN,and combines the temporal features extracted by gated recurrent unit GRU to realize network traffic prediction based on T-GCN..After the self-attention mechanism improves the original model,it obtains the weight of data at each moment in the network,the relationship between better fitting and previous and previous moments,and obtains an adaptive A3T-GCN model.The model can improve the prediction accuracy in the current simulation environment.Finally,the ONOS controller is used to build the system experimental environment,and the function and performance test are carried out on the Mininet simulation platform.The current network data are obtained through the intent command,which verifies the intent-driven network measurement function.Experiments are carried out on the performance of the A3T-GCN model in many aspects,and the best effect of the model is obtained by comparison,and the traffic prediction function of real-time data is realized.