摘要
针对现有算法存在的适用性不够全面、检测精准度较低的问题,优化移动网络时间隐蔽信道检测算法。通过二进制串编码,构建移动网络时间隐蔽信道结构,采用聚类算法获取多个内部元素相似的子类,将相邻子类之间的相似度抽象成两点间的模,完成较高密度区域的子类划分。利用密度聚类算法进行聚类,标记所得的聚类核个数与坐标,经过归一化处理属性分类,实现高维数据向低维空间的投影。滤除聚类属性,架构新的属性集合,并基于数据包传输的间隔时长序列,分别设计正常信道和隐蔽信道下的网络间隔时长模型,通过比较两种模型相邻子类之间余弦相似度均值等指标完成信道检测。仿真结果表明,所提算法能够有效检测主动式和被动式隐蔽信道,且检测准确率较高,具有更好的适用性。
Due to insufficient applicability and low detection accuracy, the algorithm to detect mobile network time covert channel was optimized. Firstly, the binary strings were used to construct the time covert channel in mobile network. Then, the clustering algorithm was used to obtain multiple subclasses with similar internal elements. After that, the similarity between adjacent subclasses was abstracted into a module between two points to complete the sub-class division of high-density regions. The density clustering algorithm was adopted, and then the number of clustering cores and coordinates were marked. After the normalization of attribute classification, the projection of high-dimensional data to low-dimensional space was completed. After the clustering attributes were filter out, a new set of attributes was established. Based on the interval time series of packet transmission, the models of network interval time on normal channel and covert channel were designed. Finally, the channel detection was completed by comparing the average cosine similarity between adjacent subclasses of the two models. Following conclusions can be drawn from simulation results: the proposed algorithm can effectively detect the active and passive covert channels;Meanwhile, the detection accuracy is higher than traditional method, so this method has better applicability.
作者
张博
ZHANG Bo(School of Computer Science and Technology Beijing Institute of Technology,Beijing 100081,China)
出处
《计算机仿真》
北大核心
2022年第2期180-183,211,共5页
Computer Simulation
基金
国家自然科学基金(U1636213)。
关键词
移动网络
时间隐蔽信道
检测算法
聚类
余弦相似度
Mobile network
Time covert channel
Detection algorithm
Clustering
Cosine similarity