Research On Anomaly Detection And Route Algorithm For Software Defined Network

Software Defined Network(SDN)is a new network architecture with the characteristics of centralized management and programmability.Compared with traditional networks,SDN can simplify network management and better support the dynamic control of network traffic.Security network devices can automatically detect security issues and implement solutions to prevent attacks and block the spread of threats.As network scale continues to grow and complexity increases,phenomena such as SDN device failures,malicious attacks on networks,and abnormal traffic become increasingly common and diverse.Efficient and timely detection of abnormal SDN devices and determining whether the network is experiencing abnormal conditions becomes particularly important.Applying the SDN concept to Wireless Sensor Networks(WSN),Software Defined Wireless Sensor Networks(SDWSN)emerged,solving problems such as limited node energy,computing power,and communication ability that exist in traditional WSNs.However,SDWSN still faces significant challenges such as energy consumption and network lifespan.Regarding the issues surrounding the development of SDN,the following work has been completed:(1)To address the current inadequacies of SDN architecture in defending against various attacks and resolving controller failures,an improved self-secure SDN controller architecture has been developed.Based on the basic network architecture of SDN,the overall security of the network is improved by adding a self-security management module and a route optimization module.(2)To enhance the anomaly detection and fault tolerance capabilities of SDN,an anomaly detection method based on the Byzantine Fault Tolerance(BFT)mechanism has been proposed.The effectiveness and security of this method have been theoretically analyzed and proven.Simulation results show that in an SDN environment,this detection method can quickly detect abnormal network devices,reducing the false positive and false negative rates in SDN anomaly detection.The average detection time is reduced by approximately 55.35% compared to the efficient and Byzantine fault-tolerant SDN control plane adaptive framework(MORPH)and approximately 65.45% compared to the SDN Malicious Switch Detection Prototype System(SDN-MSD).(3)An optimized algorithm based on Firefly Algorithm,Gravity Search Algorithm,and Biogeography Optimization(FGB)is designed to improve the lifespan of SDWSNs and reduce data transmission energy consumption.Energy entropy and FGB are used for cluster head selection,and a Distributed High-Efficiency Entropy Energy-Saving Cluster Routing Algorithm(DHEEC)is proposed.Experimental simulation results show that FGB performs better than other optimization algorithms on standard functions.DHEEC has about a 41.05%improvement in the number of dead nodes compared to Optimized Energy Efficient Routing Algorithm Based on Improved Firefly Clustering for Heterogeneous WSNs(IFCEER)and about a 13.89% improvement compared to Distribute Energy-Efficient Clustering with Firefly Algorithm(DEEC-FA),and improves energy utilization efficiency.(4)To timely detect network attack behaviors,an SDN-based self-security management system was developed.The system adopted a deployment scheme of Flask+Gunicorn+Nginx and achieved functions including application analysis,user analysis,abnormal traffic information analysis,and SDN speed limit configuration.It was integrated into a well-known company’s SDN product and tested by third-party detection agencies.The results showed that the system met the corresponding technical indicators in terms of security,performance,statistical capability,reliability,and functionality.