删除或更新信息,请邮件至freekaoyan#163.com(#换成@)

基于递归特性的网络应用流量行为分析

清华大学 辅仁网/2017-07-07

基于递归特性的网络应用流量行为分析
袁静1,2,王俊松1,3,李强1,陈曦1()
2. 国家计算机网络应急技术处理协调中心, 北京 100029
3. 天津医科大学 生物医药工程学院, 天津 300070
Recurrence based nonlinear analysis for network application traffic
Jing YUAN1,2,Junsong WANG1,3,Qiang LI1,Xi CHEN1()
1. Department of Automation, Tsinghua University, Beijing 100084, China
2. National Computer Network Emergency Response Technical Team Coordination Center of China, Beijing 100029, China
3. School of Biomedical Engineering, Tianjin Medical University, Tianjin 300070, China

摘要:
HTML
输出: BibTeX | EndNote (RIS)
摘要准确刻画不同网络应用流量的行为特征,是识别和控制应用流量以及保证互联网服务质量的关键。该文提出一种基于流量系统状态递归特性的分析方法,研究应用流量的内在动力学行为。针对实际网络中多类不同应用流量的时间序列,首先重构流量序列的高维相空间,然后分析应用流量系统状态运动轨迹的递归特性,揭示其各自固有的内在行为。实验结果表明,流量的非线性动力学特征能够准确地刻画各类网络应用流量的行为,并且不随网络规模或网络协议版本的改变而发生变化。因此,流量的非线性动力学特征有助于提高互联网应用流量识别与控制的性能。

关键词 网络应用流量,相空间,递归特性,动力学特征
Abstract:Accurate characterization of the traffic from different network applications plays an important role in traffic classifications to guarantee the quality of service of Internet traffic. The behavior of various network application traffic was analyzed based on the recurrence properties of the network traffic. A high-dimensional phase space is constructed for the traffic time series and then recurrences in the traffic state trajectory are analyzed to identify the intrinsic characteristics of the application traffic. Analyses show that the nonlinear dynamic features can accurately characterize application traffic behavior and that these features are independent of the network scale and Internet protocol version. Therefore, the nonlinear dynamics of application traffic can be used to improve network traffic classification.

Key wordsnetwork application trafficphase spacerecurrence propertydynamic feature
收稿日期: 2013-09-11 出版日期: 2015-04-17
基金资助:国家自然科学基金资助项目(61203039);高等学校学科创新引智计划“111计划”项目(B06002)
引用本文:
袁静,王俊松,李强,陈曦. 基于递归特性的网络应用流量行为分析[J]. 清华大学学报(自然科学版), 2014, 54(4): 515-521.
Jing YUAN,Junsong WANG,Qiang LI,Xi CHEN. Recurrence based nonlinear analysis for network application traffic. Journal of Tsinghua University(Science and Technology), 2014, 54(4): 515-521.
链接本文:
http://jst.tsinghuajournals.com/CN/ http://jst.tsinghuajournals.com/CN/Y2014/V54/I4/515


图表:
不同网络应用协议流量序列互信息与τ的关系
不同网络应用协议流量序列虚假近邻点比例与m的关系
HTTP和FTP流量的递归图
类型 采集时间 包数 流数 IP数
h 109 107 105
IPv4 24 10.7 24.6 63.3
IPv6 24 15.6 7.7 2.4


实验数据集说明
应用协议类型 端口 描述
HTTP 80、 443 Web应用
FTP 20、 21 文件传输协议
Email 25、 993 邮件传输协议
DNS 53 (UDP) 域名解析
BitTorrent 6 881 P2P应用
QQ 8 000、 4 000(UDP) 即时聊天应用


网络应用协议类型
不同网络应用协议流量的递归图
应用类型 RR/% DET/% ENT
HTTP 8.36 27.34 1.88
DNS 3.93 12.08 1.52
FTP 34.62 89.34 3.08
Email 11.77 46.75 1.97
BitTorrent 5.38 20.94 1.44
QQ 9.92 34.19 2.01


不同网络应用流量的RQA结果
不同网络应用协议流量的递归图
不同网络应用协议流量的RAQ
不同网络应用协议流量的递归图
不同网络应用协议流量的RAQ


参考文献:
[1] Ohira T, Sawatari R. Phase Transition in a computer network traffic model [J]. Physics Review E, 1998, 58(1): 193-195.
[2] Takayasu M, Takayasu H, Fukuda K. Dynamic phase transition observed in the Internet traffic flow[J]. Physica A, 2000, 277(1): 248-255.
[3] Bigdeli N, Haeri M, Choobkar S, et al.Characterization of complex behaviors of TCP/RED computer networks based on nonlinear time series analysis methods[J]. Physica D, 2007, 233(2): 138-150.
[4] LI Qiang, QIN Tao, GUAN Xionghong, et al.Empirical analysis and comparison of IPv4-IPv6 traffic: A case study on the campus network [C]// Proceedings of the 18th IEEE International Conference on Networks. Singapore: IEEE press, 2012: 395-399.
[5] Bujlow T, Riaz T, Pedersen J M. A method for classification of network traffic based on C5.0 machine learning algorithm [C]// Proceedings of International Conference on Computing, Networking and Communications. Maui, HI: IEEE press, 2012: 237-241.
[6] Grossglauser M, Bolot J C. On the relevance of long-range dependence in network traffic[J]. IEEE/ACM Transactions on Networking, 1999, 7(5): 629-640.
[7] Willinger W, Paxson V, Taqqu M S. Self-Similarity and Heavy Tails: Structural Modeling of Network Traffic[M]. Boston, MA: Springer press, 1998.
[8] Masugi M. Recurrence plot-based approach to the analysis of IP-network traffic in terms of assessing non-stationary transitions over time[J]. IEEE Transactions on Circuites and Systems, 2006 53(10): 2318-2326.
[9] WANG Junsong, YUAN Jing, LI Qiang, et al.Correlation dimension based nonlinear analysis of network traffics with different application protocols[J]. Chinese Physics B, 2011, 20(5): 050506.
[10] Kantz H, Schreiber T. Nonlinear Time Series Analysis [M]. 2nd ED. New York, NY: Cambridge University Press, 2004.
[11] Eckmann J P, Kamphorst S O, Ruelle D. Recurrence plots of dynamical systems[J]. Europhysics Letters, 1987, 4(9): 973-977.
[12] Marwan N, Romano M C, Thiel M, et al.Recurrence plots for the analysis of complex systems[J]. Physics Reports, 2007, 438(5): 237-329.
[13] Zbilut J P, Webber C L. Embedding and delays as derived from quantification of recurrence plots[J]. Physics Letter A, 1992, 171(3): 973-977.


相关文章:
No related articles found!

相关话题/网络 序列 空间 实验 控制