基于马尔可夫链改进ANN的SQL server数据库攻击入侵预警方法  

Improved ANN Method of SQL Server Database Attack Intrusion Warning Based on Markov Chain

作　　者：韩楠楠 严风 HAN Nan-nan;YAN Feng(Henan University of Science and Technology,SanMenXia Polytechni,Sanmenxia 472000,China)

机构地区：[1]河南科技大学应用工程学院,三门峡职业技术学院,河南三门峡472000

出　　处：《电脑与电信》2025年第1期19-22,共4页Computer & Telecommunication

基　　金：2022年三门峡职业技术学院教学改革研究项目“基于《数据库应用基础》课程的新时代高校思政课话语体系创新研究”,项目编号:SJG-2022-065。

摘　　要：阈值静态僵化性障碍导致了基于固定阈值的预警方法在适应数据库运行状态变化方面存在局限,进而使得预警准确性降低。为此,提出基于马尔可夫链改进人工神经网络的SQL server数据库攻击入侵预警方法。利用SQL Server的Profiler工具收集详尽的数据库操作历史数据,并提取独立性特征,形成高质量的数据源。设计三层结构的ANN模型,并引入马尔可夫链和径向基函数神经网络改进模型,从而建立了能够揭示攻击强度动态变化并实现对SQL Server数据库攻击入侵精准预警的模型。基于上述模型,实施实时预警机制,通过动态调整预警阈值和即时警报,及时发现并响应潜在的数据库安全威胁。实验结果表明,相比于传统方法,该方法能够更全面精准地预警SQL Server数据库攻击,提高了预警的准确性。The static rigidity of threshold leads to the limitation of the early warning method based on fixed threshold in adapting to the change of database operating state,which reduces the accuracy of early warning.Therefore,this paper presents a method of SQL server database attack intrusion warning based on Markov chain improved artificial neural network.The Profiler tool of SQL Server is used to collect detailed database operation history data and extract independence characteristics to form high-quality data sources.A three-layer ANN model is designed,and Markov chain and radial basis function neural network are introduced to improve the model,so as to establish a model that can reveal the dynamic change of attack intensity and realize accurate early warning of SQL Server database attacks.Based on the above model,a real-time early warning mechanism is implemented to detect and respond to potential database security threats by dynamically adjusting the early warning threshold and instant alarm.The experimental results show that compared with the traditional methods,the proposed method can predict the SQL Server database attack more comprehensively and accurately,and improve the accuracy of early warning.

关 键 词：人工神经网络 SQL server数据库 攻击入侵预警 预警模型 马尔尔科夫链 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]