[1]朱光明,冯家伟,卢梓杰,等.因果图表征的网络攻击数据集构建[J].计算机技术与发展,2024,34(04):124-131.[doi:10. 3969 / j. issn. 1673-629X. 2024. 04. 019]
　ZHU Guang-ming,FENG Jia-wei,LU Zi-jie,et al.Network Attack Dataset Construction Using Causal Graph[J].,2024,34(04):124-131.[doi:10. 3969 / j. issn. 1673-629X. 2024. 04. 019]

点击复制

因果图表征的网络攻击数据集构建()

分享到：

《计算机技术与发展》[ISSN:1006-6977/CN:61-1281/TN]

卷:

34

期数:

2024年04期

页码:

124-131

栏目:

网络空间安全

出版日期:

2024-04-10

文章信息/Info

Title:

Network Attack Dataset Construction Using Causal Graph

文章编号:

1673-629X(2024)04-0124-08

作者:

朱光明¹ ; 冯家伟² ; 卢梓杰¹ ; 张向东² ; 张锋军³ ; 牛作元³ ; 张 亮¹

1. 西安电子科技大学 计算机科学与技术学院,陕西 西安 710071

2. 西安电子科技大学 通信工程学院,陕西 西安 710071

3. 中国电子科技集团公司第三十研究所,四川 成都 610041

Author(s):

ZHU Guang-ming1 ; FENG Jia-wei2 ; LU Zi-jie1 ; ZHANG Xiang-dong2 ; ZHANG Feng-jun3 ; NIU Zuo-yuan3 ; ZHANG Liang1

1. School of Computer Science and Technology,Xidian University,Xi’an 710071,China;

2. School of Communication Engineering,Xidian University,Xi’an 710071,China;

3. The 30th Research Institute of China Electronics Technology Group Corporation,Chengdu 610041,China

关键词:

网络安全; 因果图; 高级可持续威胁攻击; 攻击上下文

Keywords:

network security; causal graph; advanced persistent threat attack; attack context

分类号:

TP391

DOI:

10. 3969 / j. issn. 1673-629X. 2024. 04. 019

摘要:

高级可持续威胁攻击因其多阶段可持续的特性,已经成为现阶段网络攻击的主要形式。 针对此类型攻击的检测、预测研究,不可避免地需要相关数据集的支撑。 在构建数据集时,往往需要真实的网络与主机数据。 但出于隐私与安全的考虑,很少有满足要求的开源数据集。 现有的数据集也往往只提供原始的网络流和日志数据,对长时攻击上下文解析的缺乏导致单纯地利用神经网络进行数据包的恶性甄别和预测的实用性不足。 为了解决这些问题,该文基于网络环境的真实攻击过程数据,构建并公布了一个基于因果图的网络攻击数据集。 与传统的原始网络流和日志数据集相比,该数据集充分挖掘了攻击上下文中的因果关系,可以跨长时域对高级可持续威胁攻击进行建模,方便研究人员进行攻击检测与预测的研究。 该数据集开源在 https: / / github. com / GuangmingZhu / CausalGraphAPTDataset 上。

Abstract:

Advanced persistent threat attack has become the main form of network attack because of its multi - stage sustainablecharacteristics. Datasets are necessary for researches?

on the detection and prediction of this kind of attack. Real network and host data aresuperior when constructing datasets. However,few publicly available datasets can meet the requirements,due to the privacy and securityissues. The available datasets often supply original network flows and system logs,but the absence of analysis on the long-term attackcontext results in that a straightforward using of deep neural networks to detect and predict malicious packets is not practical enough. Inorder to overcome these problems,

a causal graph based network attack dataset is constructed and released,based on the real attack data ofa network scene. Compared with the other datasets supplying original network flows and system logs simply,such dataset explores thecausality of attach context deeply and can model the long-term advanced persistent threat attack. This makes the dataset more applicablefor attack detection and prediction. The dataset is released at https: / / github. com / GuangmingZhu / CausalGraphAPTDataset.

常用功能