《计算机应用研究》|Application Research of Computers

一种概率保证的TCP单包流过滤机制

Probability-guaranteed filtering mechanism for TCP single-packet flow

免费全文下载 (已被下载 次)  
获取PDF全文
作者 张强伟,张孝国,张翔
机构 1.西南交通大学 信息网络中心,成都 610031;2.东南大学 计算机科学与工程学院,南京 211189;3.河南科技大学 信息工程学院,河南 洛阳 471023
统计 摘要被查看 次,已被下载
文章编号 1001-3695(2014)11-3413-04
DOI 10.3969/j.issn.1001-3695.2014.11.048
摘要 针对TCP流量中高比例单包流影响组流效率问题进行了研究,基于网络流特性和TCP规范提出了一种概率保证的TCP单包流过滤机制,在CERNET主干节点跨度两年的大量Traces上进行了实验。结果表明,该策略以极少的精度损失获取了大幅度的效率提升,与现有策略相比,性能和稳定性都有显著提升。
关键词 流终止策略;单包流;概率保证;优化;过滤;传输控制协议
基金项目 国家“973”计划资助项目(2009CB320505)
国家科技攻关计划资助项目(2008BAH37B04)
本文URL http://www.arocmag.com/article/01-2014-11-048.html
英文标题 Probability-guaranteed filtering mechanism for TCP single-packet flow
作者英文名 ZHANG Qiang-wei, ZHANG Xiao-guo, ZHANG Xiang
机构英文名 1. Information & Network Center, Southwest Jiaotong University, Chengdu 610031, China; 2. School of Computer Science & Engineering, Southeast University, Nanjing 211189, China; 3. College of Information Engineering, Henan University of Science & Technology, Luoyang Henan 471023, China
英文摘要 A large number of single-packet flows reduce the efficiency of flow identification. It is important to explore a filtering mechanism for single-packet flows so as to improve the efficiency of flow identification. This paper proposed a probability-guaranteed filtering mechanism for TCP single-packet flows based on flow characteristics and TCP protocol specifications. The results of a lot of experiments based on Traces collected from a main channel in the CERNET and spanned two years show this mechanism can improve the efficiency of flow identification substantially at the cost of little loss of accuracy. Relative to the existing filtering strategies, this mechanism improves the performance and stability greatly.
英文关键词 flow termination strategy; single-packet flow; probability-guaranteed; optimization; filtering; TCP
参考文献 查看稿件参考文献
  [1] LI Bing-dong, SPRINGER J, BEBIS G, et al. A survey of network 〓〓flow applications[J] . Journal of Network and Computer Applications, 2013, 36(2):567-581.
[2] STROHMEIER F, DORFINGER P, TRAMMELL B. Network performance evaluation based on flow data[C] // Proc of the 7th International Wireless Communications and Mobile Computing Conference. Washington DC:IEEE Computer Society, 2011:1585-1589.
[3] DAINOTTI A, PESCAPE A, CLAFFY K C. Issues and future directions in traffic classification[J] . IEEE Network, 2012, 26(1):35 -40.
[4] WU Xiao-fei, YU Ke, WANG Xin. On the growth of Internet application flows:a complex network perspective[C] // Proc of IEEE INFOCOM. Piscataway:IEEE Press, 2011:2096-2104.
[5] CLAFFY K C, BRAUN H W, POLYZOS G C. A parameterizable methodology for Internet traffic flow profiling[J] . IEEE Journal on Selected Areas in Communications, 1995, 13(8):1481-1494.
[6] RYU B, CHENEY D, BRAUN H W. Internet flow characterization:adaptive timeout strategy and statistical modeling[C] // Proc of Workshop on Passive and Active Measurement. Berlin:Springer-Verlag, 2001:95-105.
[7] WANG Jun-feng, LI Lei, SUN Fu-chun, et al. A probability-guaranteed adaptive timeout algorithm for high-speed network flow detection[J] . Computer Networks, 2005, 48(2):215-233.
[8] Cisco Systems. NetFlow services solutions guide[EB/OL] . (2007-01-22)[2013-10-13] . http://www. cisco. com/en/US/docs/ios/solutions_docs/netflow/nfwhite. pdf.
[9] 周明中, 龚俭, 丁伟. 网络流超时策略研究[J] . 通信学报, 2005, 26(4):88-93.
[10] CAI Jing, ZHANG Zhi-bin, ZHANG Peng, et al. An adaptive timeout strategy for UDP flows using SVMs[C] // Proc of Parallel and Distributed Computing, Applications and Technologies. Washington DC:IEEE Computer Society, 2010:118-127.
[11] ZHANG Xiao-guo, DING Wei. Comparative research on Internet flows characteristics[C] // Proc of the 3rd International Conference on Networking and Distributed Computing. Washington DC:IEEE Computer Society, 2012:114-118.
[12] 朱海婷, 丁伟, 缪丽华. UDP流量对TCP往返延迟的影响[J] . 通信学报, 2013, 34(1):19-29.
[13] MITCHELL T M. Machine learning[M] . Beijing:China Machine Press, 2003:96-99.
收稿日期 2013/10/13
修回日期 2013/11/28
页码 3413-3416
中图分类号 TP393.09
文献标志码 A