大规模通信网络涉密信息安全动态预警方法研究  

Research on Dynamic Early Warning Method of Classified Information Security in Large-scale Communication Network

作　　者：刘亚军[1] 沈国良 LIU Ya-jun;SHEN Guo-liang(First Affiliated Hospital of Soochow University,Suzhou Jiangsu 215006,China;Center of Information Development and Management,Soochow University,Suzhou Jiangsu 215006,China)

机构地区：[1]苏州大学附属第一医院,江苏苏州215006 [2]苏州大学信息化建设与管理中心,江苏苏州215006

出　　处：《计算机仿真》2023年第12期470-475,共6页Computer Simulation

基　　金：江苏省通品牌专业建设项目(Z215015002)。

摘　　要：为了提高大规模涉密信息在网络传输时的效率与安全,预防存储涉密信息的被入侵修改,提出一种基于双分类算法与双聚类算法融合的涉密信息安全动态预警方法。方法首先构建基于模板匹配分类与SVM分类双分类器引擎,对用户对象的行为进行分类分析;当用户对象的行为异常时,针对不同的入侵行为提出不同的安全预警策略;当用户对象的行为正常时为,通过融合SOFM与K-Means聚类算法,将涉密信息进行聚类处理,然后采用不同参数的SM2曲线加密算法对不同类的涉密信息进行上行传输,最后通过密文解密还原的方式,将涉密信息下行给用户对象。仿真结果表明,较其它静态预警方法相比,提出的动态预警方法在传输涉密信息的保护上,当并发用户数量为500时,其涉密信息传输时间很短,仅为5.21s;在存储涉密信息的预警上,丢包率均值最低仅为3.22。因此,本文提出的方法在涉密信息存储与传输上,效率更快、动态性更好,安全性更高。In order to improve the efficiency and security of large-scale classified information transmission in the network and prevent the storage of classified information from being modified by the intrusion,this paper proposes a dynamic early warning method for classified information security based on the fusion of double classification algorithm and double clustering algorithm.In that method,a double classify engine based on template match classification and SVM classification is constructed to classify and analyze the behavior of a user object,and when the behavior of the user object is abnormal,different security early war strategies are proposed for different intrusion behaviors.When the behavior of the user object is normal,the classified information is clustered by fusing the SOFM and K-Means cluste⁃ring algorithms,then the SM2 curve encryption algorithm with different parameters is used to transmit different types of classified information,and finally,the classified information is sent to the user object through the way of ciphertext decryption and restoration.The simulation results show that,compared with other static warning methods,the dynamic warning method proposed in this paper has a very short transmission time of 5.21 s for the transmission of classified information when the number of concurrent users is 500,and the minimum average packet loss rate is only 3.22 for the warning of storing classified information.Therefore,the method proposed in this paper is more efficient,more dynamic and more secure in the storage and transmission of confidential information.

关 键 词：网络安全 算法融合 动态预警 

分 类 号：TN915.08[电子电信—通信与信息系统]