摘要
将态势感知的先进思想引入网络传输领域,以空间流量聚类为基本思想,建立网络传输态势感知(NTSA)模型;围绕模型关键技术,依据信息增益和互信息的等价性执行态势因子选择,提出了一种面向传输模式划分的高维数据流聚类算法,并且基于图论进行拓扑重要性分析;设计并且实现了NTSA原型系统。基于真实数据集的实验验证了系统的时效性、准确性以及可扩展性。
The advanced ideas of situation awareness were introduced to network transmission and NTSA(network transmission situation awareness) model was established based on spatial traffic clustering.Around the key technologies of the model,situation factors were selected according to information gain and mutual information;a high-dimensional data stream clustering algorithm for transmission pattern partition as well as a topology importance analysis method of network element based on graph theory were proposed;furthermore,a NTSA prototype system was designed and imple-mented.The experiment results on real datasets demonstrate the efficiency,effectiveness and scalability.
出处
《通信学报》
EI
CSCD
北大核心
2010年第9期54-63,共10页
Journal on Communications
基金
国家重点基础研究发展计划("973"计划)基金资助项目(2009CB320503)
国家高技术研究发展计划("863"计划)基金资助项目(2008AA01A325)~~
关键词
计算机体系结构
NTSA
模型
空间流量分析
聚类
特征选择
图论
computer architecture
NTSA
model
spatial traffic analysis
clustering
feature selection
graph theory