Research On Security Technology Of Cyber Physical System Based On Software-defined Network

With the development and standardization of embedded systems and sensors,Cyber Physical System(CPS)marks the transition from traditional systems to intelligent systems,and has become the core and foundation of industrial control systems.It is used in many fields such as production and life.The extensive and successful application of NET provides users with safe,reliable and stable services,which has become a popular trend in personal and industry development.Compared with the traditional network,CPS faces different challenges.The large scale of the underlying physical equipment of CPS makes management more difficult.The closed-loop interaction characteristics also require that the distance between data and calculation is better.Its application scenarios make the system under attack.Not only will it cause information loss,it may also cause serious consequences that cannot be predicted.Therefore,the safety of CPS has received more and more attention in recent years.Traditional network security algorithms have limitations such as low accuracy,low real-time performance,and complex parameter settings,and cannot fully meet the needs of CPS.Therefore,on the basis of searching,analyzing and summarizing relevant documents,this article aims at the network security requirements of CPS and uses the idea of Software Defined Network(SDN)to obtain better network management capabilities,and proposes a SDN for CPS Safety detection architecture,and propose a safety detection algorithm.The specific work done in this article includes:(1)Research on SDCPS architecture.Analyze the security problems and solutions faced by CPS,study the concepts,characteristics and connotations of CPS and SDN,use SDN-based security detection systems to improve the security of CPS,summarize and explain the SDCPS architecture,and improve CPS security in combination with SDN The superiority of SDCPS was discussed and analyzed in many aspects,and finally used AADL to model and verify the general structure of SDCPS.(2)A network security algorithm is proposed for the SDCPS architecture.After analyzing the requirements,select network security scheme based on Brainstorm Optimization algorithm(BSO)and Extreme Learning Machine(ELM).Aiming at the problem that ELM is susceptible to outliers and BSO algorithm is easy to fall into local optimality,the weight threshold assignment method of ELM and the update and mutation strategy of BSO algorithm are improved to further improve the performance of the algorithm.The Modified Brain Storm Optimization algorithm(MBSO)and Local Distance-based Weighted Regular Extreme Learning machine(LDWRELM)algorithm are proposed The four standard test functions verify the effectiveness of the MBSO algorithm improvement,and finally a security detection model based on MBSO-LDWRELM is proposed.(3)In order to verify the effectiveness of the safety detection model,a data stream-based safety detection simulation experiment is carried out.Simulate normal traffic and abnormal traffic in the SDN simulation network environment built by Mininet and collect the traffic data in the SDN controller,filter and calculate the characteristics.Finally construct the SDN traffic dataset,and select NSL-KDD dataset.Finally,the algorithm is compared with several advanced algorithms with accuracy,detection rate,recall rate,false alarm rate and time consumption as indicators.The final results show that the performance of the proposed algorithm has improved significantly.