多层分布式大型差异数据库优化入侵检测仿真

Optimized Intrusion Detection Simulation on Multi- tier Distributed and Large Differences Database

研究多层分布式大型差异数据库入侵准确检测的问题。与传统的单层数据库不同,多层数据库的层次之间的数据存在较大的差异,分类属性的差异化较小,传统的数据库入侵检测过程只是简单考虑了层与层之间的点对点数据检测,没有考虑层与层之间的相似性,放弃了应用层次间分类对检测属性的优化,导致检测准确性不高。为了避免上述传统算法的缺陷,提出了一种基于粒子群辨别树算法的多层分布式大型差异数据库入侵检测方法。提取计算机数据库的入侵特征,并且将其作为数据库入侵检测的依据。建立粒子群辨别树,对节电进行分层处理,通过运算不同层上的数据库入侵检测的概率,从而实现多层分布式大型差异数据库的入侵检测。实验结果表明,利用上述算法进行多层分布式大型差异数据库的入侵检测,可以提高数据库的安全性,保证数据库的安全运行。

In this paper, the accurately intrusion detection problems in multi-tier distributed and large differences database were researched. Different from the traditional single-database, there exists a big difference between multi- level data in the multi-tier database, and the difference is small categorical attributes. In the traditional database in- trusion detection process, it only simply considers the point-to-point data detection between the layers, without con- sidering the similarities between the layers, and gives up the optimization for detected attribute by application level classification, resulting in accurate detection is not high. To avoid these shortcomings of traditional algorithm, this paper presented an intrusion detection method based on particle swarm identification tree algorithm for multi-tier dis- tributed and large differences database. We extracted the features of a computer database intrusion and used it as a basis for the database, then created a PSO to identify the tree, treated the point by layer, and through the probability operation of the database on the different layers in intrusion detection, achieved intrusion detection in multi-tier dis- tributed and larger differences database. The experimental results show that the algorithm applied in intrusion detec- tion of multilayer distributed and large difference database can improve the security of the database and ensure the safe operation of database.