• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文学位 >Investigating the robustness of machine learning based IDSs on denial of service attacks for 802.11 networks.
【24h】

Investigating the robustness of machine learning based IDSs on denial of service attacks for 802.11 networks.

机译：研究基于机器学习的IDS对802.11网络的拒绝服务攻击的鲁棒性。

获取原文
获取原文并翻译 | 示例
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

Recent research has shown that machine learning based Intrusion Detection Systems (IDSs) can be used to detect link layer attacks in 802.11 networks. These machine learning based detectors are not the first attempts to produce IDSs for 802.11 networks; conventional detectors like Snort-Wireless, Netstumbler and Kismet already exist. If these machine learning based detectors are to be considered viable alternatives to the conventional detectors, we would need to know more about how robust they are. Do they have any advantages over conventional detectors? Do they have any disadvantages? If so what can be done to mitigate them?; Focusing primarily on stealth attacks, cross-platform performance and performance in detecting unknown similar attacks; the objective of this research is to investigate the questions posed above. Results show that machine learning based detectors, specifically GP-based detectors, showed better survivability than conventional detectors i.e. Snort-Wireless, in the face of a stealth DoS attack and where able to detect other similar unknown Dos attacks which they were not trained on. Also using novel map address mapping techniques in the processing of our training datasets we were able to show that machine learning based IDSs can be cross-platform robust like their conventional counterparts.
机译：最近的研究表明，基于机器学习的入侵检测系统（IDS）可用于检测802.11网络中的链路层攻击。这些基于机器学习的检测器并不是为802.11网络生成IDS的第一个尝试。诸如Snort-Wireless，Netstumbler和Kismet之类的常规探测器已经存在。如果这些基于机器学习的检测器被认为是传统检测器的可行替代方案，我们将需要更多地了解它们的鲁棒性。与传统的探测器相比，它们有什么优势吗？他们有什么缺点吗？如果是这样，可以采取什么措施来减轻他们的负担？主要侧重于隐身攻击，跨平台性能以及检测未知相似攻击的性能；这项研究的目的是调查上面提出的问题。结果表明，基于机器学习的检测器（特别是基于GP的检测器）比常规检测器（即Snort-Wireless）表现出更好的生存能力，从而可以应对隐形DoS攻击，并且能够检测到其他未经训练的类似未知Dos攻击。在训练数据集的处理中还使用了新颖的地图地址映射技术，我们能够证明基于机器学习的IDS可以像传统的IDS一样具有跨平台的鲁棒性。

著录项

  • 作者

    Makanju, Adetokunbo.;

  • 作者单位

    Dalhousie University (Canada).;

  • 授予单位 Dalhousie University (Canada).;
  • 学科 Engineering Electronics and Electrical.; Computer Science.
  • 学位 M.Sc.
  • 年度 2008
  • 页码 125 p.
  • 总页数 125
  • 原文格式 PDF
  • 正文语种 eng
  • 中图分类
  • 关键词

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号