摘要
In recent years, because the number of the computers in the Internet has been increasing, it is important for network accounting, network safe management and network flow measurement to see the behavior of the activated IP flow. In the paper, we research almost 30TB traffic trace that all traffic entered into and removed from the CERNET Eastern China (north) regional network in Feb. 2001. And we analyze the active degree of IP in the regional network. The research result indicate that the statistics of in-IP flow, out-IP flow and IP flow are heavytailed and the ratio of in-IP flow and out-IP flow have the character of log-normal distribution. According to these characters, we make the model of pareto distribution and log-normal distribution and propose the way that evaluate these distribution models.
In recent years, because the number of the computers in the Internet has been increasing. it is important for network accounting, network safe management and network flow measurement to see the behavior of the activated IP flow. In the paper, we research almost 30TB traffic trace that all traffic entered into and removed from the CERNET Eastern China (north) regional network in Feb. 2001. And we analyze the active degree of IP in the regional network. The research result indicate that the statistics of in-IP flow, out-IP flow and IP flow are heavy-tailed and the ratio of in-IP flow and out-IP flow have the character of log-normal distribution. According to these characters, we make the model of pareto distribution and log-normal distribution and propose the way that evaluate these distribution models.
出处
《计算机科学》
CSCD
北大核心
2003年第4期43-46,共4页
Computer Science
基金
国家自然科学基金"90104031"
国家863课题"2001AA112060"
