基于K-means聚类算法的网络入侵监测系统设计

Design of network intrusion detection system based on K-means clustering algorithm

常规的网络入侵监测系统对于高维度数据,存在传输速率慢的问题。为此,设计基于K-means聚类算法的网络入侵监测系统。在硬件设计上,引用安全加速芯片,将其安置在内部高速总线上,通过DMA实现芯片与CPU的连接;在软件设计上,预处理监测的网络数据,将其作为算法的输入,利用K-means聚类算法使网络入侵数据执行聚类操作,以目标函数收敛为依据,输出聚类结果,根据结果判断网络入侵类型,并采取相应的措施;结合硬件设计和软件设计完成网络入侵监测系统设计。测试结果表明,在数据集相同的情况下,设计的基于K-means聚类算法的网络入侵监测系统传输速率为87.43 Mb/s,远高于常规的网络入侵监测系统,说明该系统更适合应用在实际项目中。

Conventional network intrusion monitoring system has the problem of slow transmission rate for high dimensional data.Therefore,a network intrusion detection system based on machine learning algorithm is designed.In terms of hardware design,the security acceleration chip is referred to and placed on the internal high-speed bus to connect the chip and CPU through DMA.In terms of software design,the monitored network data is preprocessed and taken as the input of the algorithm.Machine learning algorithm is used to perform clustering operation on the network intrusion data.According to the convergence of the objective function,clustering results are output.The design of network intrusion monitoring system is completed by combining hardware design and software design.The test results show that under the same data set,the designed network intrusion detection system based on machine learning algorithm has a transmission rate of 87.43 mb/s,which is much higher than the conventional network intrusion detection system,indicating that the system is more suitable for application in actual projects.