信息网络安全 ›› 2020, Vol. 20 ›› Issue (12): 28-32.doi: 10.3969/j.issn.1671-1122.2020.12.004

• 技术研究 • 上一篇    下一篇

一种自适应的异常流量检测方法

张新跃1, 胡安磊1(), 李炬嵘1, 冯燕春2   

  1. 1.中国互联网络信息中心,北京 100190
    2.国家信息技术安全研究中心,北京 100044
  • 收稿日期:2020-08-18 出版日期:2020-12-10 发布日期:2021-01-12
  • 通讯作者: 胡安磊 E-mail:huanlei@cnnic.cn
  • 作者简介:张新跃(1978—),男,云南,正高级工程师,博士,主要研究方向为DNS安全、云安全、网络安全|胡安磊(1979—),男,山东,正高级工程师,硕士,主要研究方向为DNS安全、网络安全|李炬嵘(1976—),女,新疆,高级工程师,硕士,主要研究方向为DNS安全、网络安全|冯燕春(1960—),女,北京,研究员,本科,主要研究方向为风险评估、网络安全
  • 基金资助:
    科技部重点研发(2019YFB1804501)

A Method of Adaptive Abnormal Network Traffic Detection

ZHANG Xinyue1, HU Anlei1(), LI Jurong1, FENG Yanchun2   

  1. 1. China Internet Network Information Center, Beijing 100190, China
    2. National Research Center for Information Technology Security, Beijing 100044, China
  • Received:2020-08-18 Online:2020-12-10 Published:2021-01-12
  • Contact: HU Anlei E-mail:huanlei@cnnic.cn

摘要:

文章针对DDoS异常流量攻击提出一种自适应攻击检测方法,该方法基于网络访问行为特征进行快速学习建模,再通过一个流量TOP-N排序表来实现异常流量的动态过滤。TOP-N排序表的样本模板采用自适应收敛算法来快速自学习更新,可以快速、准确地识别出异常流量和攻击行为,大大提升异常流量攻击行为检测的准确率,特别适用于慢速的应用型DDoS攻击检测和防护领域。

关键词: DDoS, TOP-N, 自适应, 训练模板

Abstract:

In this paper, we propose a new adaptive attack detection method for DDoS abnormal traffic attacks. The method is based on the characteristics of network access behavior for rapid learning modeling, and then through a traffic TOP-N ranking table to achieve dynamic filtering of abnormal traffic. The sample template of TOP-N table adopts adaptive convergence algorithm to quickly self-learning update. This method can quickly and accurately identify abnormal traffic and attack behavior, and greatly improve the accuracy of abnormal traffic attack detection. It is especially suitable for the detection and protection of slow application DDoS attacks.

Key words: DDoS, TOP-N, adaptive, training template

中图分类号: