【摘要】 通过分析2009年1月至2010年4月间采集的CAIDA自治域级数据,研究发现:网络规模、节点间连接程度和聚集性随时间有增加的趋势.依据k-核解析的概念将网络分成不同的层,具体针对网络特征值有跳跃性变化的时间点,分析该时间点各层度分布、层节点数、层连接分布和层节点度总和的变化.结果表明:此时间点低层节点数明显减少但连接概率增加,影响网络连接程度的主要因素是中间层间节点的连接变化,突变性变化使网络处于不稳定状态,网络有自我调节能力,会逐渐恢复稳定状态.利用信息熵概念引入适用于Internet的标准结构熵概念,进一步验证了拓扑中小涨落的影响积累到一定程度时,系统就会产生"巨涨落",表现为特征值突变.更多还原

【Abstract】 AS-level CAIDA data analysis conducted from January 2009 to April 2010 found that network size,degree correlations,and clustering all tend to increase with time.According to the concept of k-core,the Internet topology was divided into different cores at the AS level.Quantities of primary characteristics were analyzed,including degree distribution,node number,node correlations,and sums of node degree in each shell.Node number in lower shells decreased but connection probability increased.The primary factor affecting network connectivity was changes in sums of degree in middle shells.Catastrophe changes in the network create an unstable state,but stability is gradually restored.The concept of Internet standard structure entropy is proposed based on the concept of entropy,thus further validating that when effects of small fluctuations on the topology accumulate to a certain point,the system will produce a "giant fluctuation",causing mutation of characteristic quantities.更多还原