Research on situation assessment and forecasting in multi-SDN integrated networks

Software Defined Networking (SDN) enhances the network's programmability and control capabilities through flexible configuration and centralized management of network resources, but it also introduces security risks and challenges. Network Security Situational Awareness (NSSA) technology integrates and analyzes information from multiple sources to achieve real-time monitoring of network security, but the accuracy of assessments is still low due to most research being based on traditional networks. This paper proposes a network situation assessment model that combines BP neural networks and Cuckoo Search (CS) algorithms in an integrated network environment with multiple SDN controllers, and improves the step size factor of the CS algorithm to increase accuracy. Subsequently, for network situation prediction issues, this paper proposes a prediction algorithm based on Gated Recurrent Units (GRU) and utilizes an improved Sparrow Search Algorithm (SSA) to replace manual parameter tuning, achieving hyperparameter optimization for the GRU model, thereby making prediction results more accurate. Finally, the assessment and prediction system were built and verified, with accuracy indicating that the design has good capabilities for situation assessment and prediction.