基于CNN和SVDD的DDoS攻击检测研究
发布时间:2022-04-23 20:00
互联网技术的飞速发展已在诸多方面影响甚至改变着人类的生活方式,便利了人们的生活。虽然互联网使我们的生活更加方便,但其脆弱性和通过互联网进行通信的信息量为对手在基础架构内执行恶意活动的机会。任何连接到公共互联网甚至私人网络的主机,都会不断受到潜在攻击的威胁。网络安全已经成为企业和组织考虑的一个非常重要的因素。然而,互联网的脆弱性以及其庞大的通信信息量,使得攻击者有机会在其基础架构内进行恶意攻击,从而带来严重的后果。DDoS攻击是一种非常典型的网络攻击,DDOS攻击会在通往目标系统的路径上阻塞很多资源,例如CPU功率、带宽、内存、处理时间等等。任何DDOS防御机制的主要目标都是尽快检测DDOS攻击,并使其在尽可能靠近其来源时就被发现。卷积神经网络(CNN)在任何应用领域的应用都包括许多步骤:数据的集成和预处理,机器学习模型的训练,以及基于在训练模型进行预测和决策。当应用于各种分类问题时,基于深度学习的方法优于现有的机器学习技术。他们通过剔除神经网络的非线性,以无监督的方式降低高维数据集的特征提取维数,并将深度学习应用于各种入侵检测系统的实施。深度学习是一个强大的工具,可以提供识别安全漏洞的...
【文章页数】:62 页
【学位级别】:硕士
【文章目录】:
摘要
Abstract
1 Introduction
1.1 Introduction and background
1.2 Related works
1.3 Motivation behind the Project
1.4 Scope and overview of the document
2 DDo S Attack
2.1 DDo S Attacks Classification and Architectures
2.1.1 Classification
2.1.2 DDo S attacks architectures
2.1.3 DDo S Strategy
2.2 DDo S Defense,Detection and Mitigation
2.2.1 DDo S attacks architectures
2.2.2 DDo S Detection and Mitigation Strategies
2.3 Deep Learning Approach in DDo S detection
3 DDo S attack detection based on CNN and SVDD approach
3.1 Experiment Environment
3.2 Data Preparation and data processing
3.2.1 The Data Set
3.2.2 Methodology
3.2.3 Packets feature processing
3.3 DDo S attack defense based on Deep Learning
3.3.1 Convolutional neural network
3.3.2 Support vector data description(SVDD)
3.4 The proposed CNN-SVDD approach
3.4.1 CNN feature extraction
3.4.2 SVDD Classification
3.5 Experiment results
3.6 Summary
4 Transfer Learning on DDo S attack detection
4.1 Transfer Learning Approach
4.1.1 Transfer Learning Technique
4.1.2 NSL-KDD Dataset
4.2 Experimental setting
4.3 Experiment Result on Transfer Learning Method
4.4 Summary
5 Conclusion and Future work
5.1 Conclusion
5.2 Future Development
Acknowledgement
References
Achievement
【参考文献】:
期刊论文
[1]基于核学习的入侵检测改进方法[J]. 周泽寻,蒋芸,明利特,王明芳,谢国城,李想. 计算机工程. 2012(14)
[2]基于改进小波分析的DDoS攻击检测方法[J]. 吕良福,张加万,张丹. 计算机工程. 2010(06)
[3]DDoS攻击的全局异常相关检测方法[J]. 李宗林,胡光岷,杨丹,姚兴苗. 计算机应用. 2009(11)
[4]基于支持向量数据描述的异常检测方法[J]. 杨敏,张焕国,傅建明,罗敏. 计算机工程. 2005(03)
本文编号:3647883
【文章页数】:62 页
【学位级别】:硕士
【文章目录】:
摘要
Abstract
1 Introduction
1.1 Introduction and background
1.2 Related works
1.3 Motivation behind the Project
1.4 Scope and overview of the document
2 DDo S Attack
2.1 DDo S Attacks Classification and Architectures
2.1.1 Classification
2.1.2 DDo S attacks architectures
2.1.3 DDo S Strategy
2.2 DDo S Defense,Detection and Mitigation
2.2.1 DDo S attacks architectures
2.2.2 DDo S Detection and Mitigation Strategies
2.3 Deep Learning Approach in DDo S detection
3 DDo S attack detection based on CNN and SVDD approach
3.1 Experiment Environment
3.2 Data Preparation and data processing
3.2.1 The Data Set
3.2.2 Methodology
3.2.3 Packets feature processing
3.3 DDo S attack defense based on Deep Learning
3.3.1 Convolutional neural network
3.3.2 Support vector data description(SVDD)
3.4 The proposed CNN-SVDD approach
3.4.1 CNN feature extraction
3.4.2 SVDD Classification
3.5 Experiment results
3.6 Summary
4 Transfer Learning on DDo S attack detection
4.1 Transfer Learning Approach
4.1.1 Transfer Learning Technique
4.1.2 NSL-KDD Dataset
4.2 Experimental setting
4.3 Experiment Result on Transfer Learning Method
4.4 Summary
5 Conclusion and Future work
5.1 Conclusion
5.2 Future Development
Acknowledgement
References
Achievement
【参考文献】:
期刊论文
[1]基于核学习的入侵检测改进方法[J]. 周泽寻,蒋芸,明利特,王明芳,谢国城,李想. 计算机工程. 2012(14)
[2]基于改进小波分析的DDoS攻击检测方法[J]. 吕良福,张加万,张丹. 计算机工程. 2010(06)
[3]DDoS攻击的全局异常相关检测方法[J]. 李宗林,胡光岷,杨丹,姚兴苗. 计算机应用. 2009(11)
[4]基于支持向量数据描述的异常检测方法[J]. 杨敏,张焕国,傅建明,罗敏. 计算机工程. 2005(03)
本文编号:3647883
本文链接:https://www.wllwen.com/kejilunwen/zidonghuakongzhilunwen/3647883.html