2025/07/15 by Yuting Zhu, Yangming Guo, Chilian Chen +3
Computer Science · Engineering · #Software-Defined Networks and 5G #Smart Grid Security and Resilience #Network Security and Intrusion Detection
paper · doi:10.1109/taes.2025.3589354
Abstract: The increasing demand for global seamless coverage and information sharing has propelled the rapid advancement of Space-Air-Ground Integrated Networks (SAGIN). Traditional network architectures are increasingly inadequate to satisfy the complex requirements of large-scale, highly dynamic environments. The multi-controller collaborative architecture of SAGIN, underpinned by Software-Defined Networking (SDN), addresses the load imbalance among SDN controllers induced by dynamic satellite network topologies and uneven user service demands. A phased SDN controller deployment and allocation strategy is proposed to overcome these technical challenges. Initially, the strategy refines the DPC-K-means algorithm by incorporating the K-nearest neighbors algorithm to assist in determining clustering centers and boundaries, thereby optimizing local density and relative distance to enhance clustering accuracy on complex datasets. In addition, the integration of flexible buffer zone technology ensures balanced load distribution among satellite controllers and facilitates appropriate partitioning. Subsequently, the Improved Snow Ablation Optimization (ISAO) algorithm is augmented by introducing a composite chaotic system algorithm and a Levy flight-based optimization strategy. This enhancement increases the diversity and flexibility of sample initialization, improving both the efficiency and accuracy in identifying optimal solutions and achieving better controller allocation in satellite networks. Simulation results demonstrate that the proposed controller allocation strategy effectively balances load, enhances the reliability and load-balancing capabilities of the SDN-based SAGIN, and improves the overall objective function value by 42%, time efficiency by 28.5%, while reducing the average transmission delay by 19%. This strategy more effectively meets the demands for multi-domain collaborative communication, transmission, and management in complex multi-platform network scenarios, providing robust technical support for SAGIN applications.