【作者】 戴宣； 丁伟；

【Author】 DAI Xuan, DING Wei (Dept of Computer Science & Engineering, Southeast University, Nanjing 210096) (Key Lab of Computer Network Technology in Jiangsu Province)

【机构】 东南大学计算机系； 江苏省计算机网络技术重点实验室；

【摘要】 本文提出了一种基于索引的数据流存储模型,它可以在不存储报文所有信息的前提下,完成面向数据流的有关测度和特性的计算,而这些参数如果采用传统方式需要在重复读取数据流信息的条件下才能获得。文章首先介绍了模型的原理和实现方案,然后完成了与传统方式的比较,最后是基于实际环境的测试结果。这些结果可以用于与网络行为学有关的研究。更多还原

【Abstract】 An index-based storage model for flow is proposed. With the model, the calculation of data flow metrics and its characteristic can be completed, instead of storing the whole information of packets. In traditional way, all the result should be obtained by re-extracting data flow information again and again. First, the paper gives the description on principle and scheme of the model; then, a comparison between the model and traditional way is made; finally, a trace is analyzed with this model. Some interesting characters appeared, which can be used for network behavior research.更多还原