摘要
采用当前方法分配网络中的密钥时,不能精准的检测出网络中存在的异常数据,分配密钥所用的时间较长,存在异常数据检测误报率高和密钥分配效率低的问题。提出分簇结构跨层连接网络的随机密钥预分配方法,通过主成分分析法和小波变换方法构建网络正常流量模型,在网络正常流量模型的基础上采用EWMA控制图和Shewart控制图分析网络中存在的残余流量,根据分析结果检测网络中存在的异常数据。去除网络中存在的异常数据后,选择簇首实现分簇,通过基站将密钥分配给簇首、簇首将密钥分配给簇内节点、节点在相同簇内的通信以及不同簇中节点之间的通信实现密钥的预分配。仿真结果表明,所提方法检测网络异常数据的误报率低、密钥分配效率高。
The current methods cannot accurately detect the abnormal data in network during the allocation of keys. Due to high false alarm rate of abnormal data detection and low efficiency of key allocation, this article puts forward a method for pre-allocate the random key in clustered cross-layer connection network with clustering structure. The normal network traffic model was constructed by principal component analysis and wavelet transform. Based on the normal network traffic model, EWMA control chart and Shewart control chart were used to analyze the residual traffic in network. According to the analysis results, the abnormal data in network were detected. After removing the abnormal data in network, the cluster head was selected to realize the clustering. Through the base station, the keys were allocated to the cluster head, and then the cluster head allocated the keys to the nodes within the cluster head. Based on the communication between nodes in the same cluster or different clusters, the key pre-distribution was achieved. Simulation results show that the proposed method has low false alarm rate and high efficiency of key allocation in detecting network abnormal data.
作者
常国锋
CHANG Guo-feng(Xinxiang college,College of Computer and Information Engineering,Henan Xinxiang 453003,China)
出处
《计算机仿真》
北大核心
2019年第11期260-263,267,共5页
Computer Simulation
关键词
分簇结构
随机密钥
分配方法
主成分分析
Clustering structure
Random key
Allocation method
Principal component analysis(PCA)
作者简介
常国锋(1978-),男(汉族),河南濮阳人,硕士,副教授,研究方向:计算机网络,网络安全。
