基于多尺度卷积神经网络的DDoS攻击检测方法

近年来,网络安全面临着日益严峻的挑战,其中分布式拒绝服务(DDoS)攻击是网络威胁中的一种常见形式。为了应对这一挑战,提出了一种基于多尺度卷积神经网络(MSCNN)的DDoS攻击检测方法。在CICDDoS2019day1数据集训练模型,CICDDoS2019day2数据集测试模型检测性能。通过利用MSCNN对网络流量进行预测和分类,能够有效识别DDoS攻击并减少误报率。实验表明,MSCNN方法在准确性、召回率、F1得分性能指标上优于SVM、DNN、CNN、LSTM和GRU。

人工智能在预测和防御DDoS攻击中的应用

针对分布式拒绝服务(DDoS)攻击在网络安全领域造成的严峻挑战,文中提出了一种基于人工智能的DDoS攻击预测模型,并详细讨论了模型的构建、数据集与特征选择、模型训练与验证过程。同时,探讨了人工智能技术在DDoS攻击实时监控、异常检测、自动化响应及模型优化调整方面的应用。结果表明,利用人工智能技术可以有效提高针对DDoS攻击的预测准确性和防御效率,为网络安全提供新的解决方案。

Blue Screen of Death Observed for Microsoft Windows Server 2012 R2 under DDoS Security Attack

Microsoft server Operating Systems are considered to have in-built, host based security features that should provide some protection against Distributed Denial of Service (DDoS) attacks. In this paper, we presented results of experiments that were conducted to test the security capability of the latest server Operating System from Microsoft Inc., namely Windows Server 2012 R2. Experiments were designed to evaluate its in-built security features in defending against a common Distributed Denial of Service (DDoS) attack, namely the TCP-SYN based DDoS attack. Surprisingly, it was found that the Windows Server 2012 R2 OS lacked sufficient host-based protection and was found to be unable to defend against even a medium intensity3.1 Gbps-magnitude of TCP-SYN attack traffic. The server was found to crash within minutes after displaying a Blue Screen of Death (BSoD) under such security attacks.

A Model Training Method for DDoS Detection Using CTGAN under 5GC Traffic

With the commercialization of 5th-generation mobile communications(5G)networks,a large-scale internet of things(IoT)environment is being built.Security is becoming increasingly crucial in 5G network environments due to the growing risk of various distributed denial of service(DDoS)attacks across vast IoT devices.Recently,research on automated intrusion detection using machine learning(ML)for 5G environments has been actively conducted.However,5G traffic has insufficient data due to privacy protection problems and imbalance problems with significantly fewer attack data.If this data is used to train an ML model,it will likely suffer from generalization errors due to not training enough different features on the attack data.Therefore,this paper aims to study a training method to mitigate the generalization error problem of the ML model that classifies IoT DDoS attacks even under conditions of insufficient and imbalanced 5G traffic.We built a 5G testbed to construct a 5G dataset for training to solve the problem of insufficient data.To solve the imbalance problem,synthetic minority oversampling technique(SMOTE)and generative adversarial network(GAN)-based conditional tabular GAN(CTGAN)of data augmentation were used.The performance of the trained ML models was compared and meaningfully analyzed regarding the generalization error problem.The experimental results showed that CTGAN decreased the accuracy and f1-score compared to the Baseline.Still,regarding the generalization error,the difference between the validation and test results was reduced by at least 1.7 and up to 22.88 times,indicating an improvement in the problem.This result suggests that the ML model training method that utilizes CTGANs to augment attack data for training data in the 5G environment mitigates the generalization error problem.

SDN环境中基于Bi-LSTM的DDoS攻击检测方案

针对5G物联网环境中海量接入设备带来的DDoS攻击威胁,同时考虑到软件定义网络SDN对5G物联网的适用性,提出了一种在SDN环境中利用长短期记忆LSTM网络检测DDoS攻击的方案,以提高对DDoS攻击检测的准确性。并采用分治算法思想,提出了一种轻量级分布式边缘计算架构OCM,在物联网中的空闲边缘节点部署基于Bi-LSTM的轻量级神经网络完成检测任务,在保证准确性的同时,增加了检测的灵活性。在ISCX2012数据集上评估了所提方案的有效性和可行性。实验结果表明,所提方案能够准确检测DDoS攻击并有效缓解DDoS攻击。

DDoS Attack in Software Defined Networks: A Survey

Distributed Denial of Service(DDoS) attacks have been one of the most destructive threats to Internet security. By decoupling the network control and data plane, software defined networking(SDN) offers a flexible network management paradigm to solve DDoS attack in traditional networks. However, the centralized nature of SDN is also a potential vulnerability for DDo S attack. In this paper, we first provide some SDN-supported mechanisms against DDoS attack in traditional networks. A systematic review of various SDN-self DDo S threats are then presented as well as the existing literatures on quickly DDoS detection and defense in SDN. Finally, some promising research directions in this field are introduced.

Effectiveness of Built-in Security Protection of Microsoft’s Windows Server 2003 against TCP SYN Based DDoS Attacks

Recent DDoS attacks against several web sites operated by SONY Playstation caused wide spread outage for several days, and loss of user account information. DDoS attacks by WikiLeaks supporters against VISA, MasterCard, and Paypal servers made headline news globally. These DDoS attack floods are known to crash, or reduce the performance of web based applications, and reduce the number of legitimate client connections/sec. TCP SYN flood is one of the common DDoS attack, and latest operating systems have some form of protection against this attack to prevent the attack in reducing the performance of web applications, and user connections. In this paper, we evaluated the performance of the TCP-SYN attack protection provided in Microsoft’s windows server 2003. It is found that the SYN attack protection provided by the server is effective in preventing attacks only at lower loads of SYN attack traffic, however this built-in protection is found to be not effective against high intensity of SYN attack traffic. Measurement results in this paper can help network operators understand the effectiveness of built-in protection mechanism that exists in millions of Windows server 2003 against one of the most popular DDoS attacks, namely the TCP SYN attack, and help enhance security of their network by additional means.

基于集成特征选择的DoS/DDoS攻击流量检测方法

随着网络规模的不断扩大以及复杂程度的不断增加,网络中拒绝服务(Denial of Service,DoS)攻击和分布式拒绝服务(Distributed Denial of Service,DDoS)攻击的发生频率越来越高。一般方法很难同时保证检测的实时性和准确性。针对上述问题,对网络流量中的DoS和DDoS攻击流量进行分析,提出了一种将过滤法和嵌入法结合的集成特征选择算法。首先使用过滤法中的相关系数法进行特征排序,按一定比例抽取特征序列组成特征子集。随后通过嵌入法中的随机森林算法对特征子集进行二次特征选择。最后通过决策树和随机森林分类器验证所提算法的分类准确率与分类效率。实验结果表明,与单一嵌入法相比,运用集成特征选择算法后,各项评价指标平均提升6%。与单一过滤法相比,仅需其特征总量的1/6即可达到同样效果。

IP Traceback技术综述

拒绝服务攻击（DoS）和分布式拒绝服务攻击（DDoS）一直是网络上使用频率最高的攻击方式。为了能够找到网络攻击的发起源以便对攻击发起人进行法律和经济上的惩罚，也为了威慑那些试图进行网络攻击的人以减少攻击发起的数量，人们开始研究IP traceback技术。本文系统地介绍了各种IP traceback技术的研究成果，对其在性能和效率上做了纵向和横向的比较，并对该技术的发展趋势做了简单的分析。

基于DDoS脚本的预防DDoS攻击策略研究

针对在中小型网站中使用硬件预防DDoS攻击造价比较昂贵的弊端,结合DDoS攻击的原理和攻击过程,提出了使用免费开源工具DDoS deflate预防DDoS攻击.从DDoS deflate的安装和配置方面做了详细的介绍,并通过web性能压力测试工具webbench模拟了对Web网站的DDoS攻击.实验结果证明了使用DDoS deflate可以有效地预防DDoS攻击,降低了Web服务架构吞吐量负载的瓶颈,不仅可以较低预防的成本,而且在一定程度上提高了网站的安全性和稳定性,具有一定的研究价值.

基于网络安全芯片的DDoS攻击识别IP核设计

分布式拒绝攻击(distributed denial of service,DDoS)作为一种传统的网络攻击方式,依旧对网络安全存在着较大的威胁.本文研究基于高性能网络安全芯片SoC+IP的构建模式,针对网络层DDoS攻击,提出了一种从硬件层面实现的DDoS攻击识别方法.根据硬件协议栈设计原理,利用逻辑电路门处理网络数据包进行拆解分析,随后对拆解后的信息进行攻击判定,将认定为攻击的数据包信息记录在攻击池中,等待主机随时读取.并通过硬件逻辑电路实现了基于该方法的DDoS攻击识别IP核(intellectual property core),IP核采用AHB总线配置寄存器的方式进行控制.在基于SV/UVM的仿真验证平台进行综合和功能性测试.实验表明,IP核满足设计要求,可实时进行DDoS攻击识别检测,有效提高高性能网络安全芯片的安全防护功能.

SDN中基于Renyi交叉熵和RMSprop算法的双层DDoS识别模型

针对软件定义网络(SDN)中存在的DDoS攻击,提出了基于Renyi交叉熵和RMSprop算法的DDoS识别模型。首先,引入计算双向流比例作为检测模块的初检方法,降低常态化监控负荷的同时,能够及时发现异常流量;然后,引入Renyi交叉熵算法作为识别模块流量特征相似性定值计算方法,有效扩大异常与正常流量数据间的信息距离,对DDoS初期小流量攻击,可以更早地识别;最后,识别模块引入RMSprop算法计算当前网络阈值,可以吸收瞬时突变,进一步提升识别准确性。实验结果表明,该模型具备时间开销低、识别成功率高的特点,可以有效地增加SDN的安全性。

一种防御DDoS攻击的软件定义安全网络机制

软件定义网络的出现为防御DDo S攻击提供了新的思路.首先,从网络体系结构角度建模分析了DDo S攻击所需的3个必要条件:连通性、隐蔽性与攻击性;然后,从破坏或限制这些必要条件的角度出发,提出了一种能够对抗DDo S攻击的软件定义安全网络机制SDSNM(software defined security networking mechanism).该机制主要在边缘SDN网络实现,同时继承了核心IP网络体系架构,具有增量部署特性.利用云计算与Chord技术设计实现了原型系统,基于原型系统的测量结果表明,SDSNM具有很好的扩展性和可用性.

面向云环境内部DDoS攻击检测的博弈论优化

结合传统基于虚拟机内省(virtual machine introspection-based,VMI)和基于网络(networkbased)入侵检测系统(intrusion detection system,IDS)的特点,提出一种部署在云服务器集群内部的协同入侵检测系统(virtual machine introspection&network-based IDS,VMI-N-IDS)来抵御云环境内部分布式拒绝服务攻击(distributed denial of service,DDoS)攻击威胁,比如"云滴冻结"攻击.将入侵检测系统和攻击者看作是博弈的双方,提出一种针对云服务器集群内部DDoS攻击与检测的博弈论模型;分别给出博弈双方的效用函数,并证明了该模型子博弈精炼纳什均衡;给出了权衡误报率和恶意软件规模控制的最佳防御策略,解决了动态调整云环境内部入侵检测策略的问题.实验表明,VMI-N-IDS能够有效抵御云环境内部DDoS攻击威胁.

DDoS攻击原理及对策研究

先通过比较DoS与DDoS的概念 ,引入了DDoS攻击的原理 ,并介绍了DDoS攻击的常用工具 ,接着重点对DDoS攻击的对策进行了比较详细的讨论 ,讨论了DDoS攻击的监测 ,以及如何防御DDoS攻击等 。

分布式DNS反射DDoS攻击检测及控制技术

分布式DNS反射DDoS攻击已经成为拒绝服务攻击的主要形式之一,传统的基于网络流量统计分析和网络流量控制技术已经不能满足防护需求。提出了基于生存时间值(TTL)智能研判的DNS反射攻击检测技术,能够准确发现伪造源IP地址分组;基于多系统融合的伪造源地址溯源阻断技术,从源头上阻断攻击流量流入网络。

DDOS攻击的分析、检测与防范技术

综述了DDOS的攻击方式的体系结构,并深入分析了DDOS的工作原理和实际的攻击工具,为进一步增强网络的安全性,根据DDOS的攻击的过程与特点,着重提出了一些DDOS的检测模型和防范DDOS攻击的技术手段。

一种用于实时追踪DDoS攻击源的分步算法

鉴于因特网出现了越来越多的DDoS攻击事件,而且这些攻击事件大多数都是利用“地址欺骗(IPSpoofing)”的攻击手段,因此DDoS攻击源追踪问题已成为网络安全研究领域的一个新方向.本文提出了一种分步追踪攻击源的新算法,其核心思想是首先由基于自治域系统(AS)的概率标记算法(ASPPM)将攻击源确定在某些AS中,然后在AS自治域范围内再使用随机数标记算法(RNPM)精确定位攻击源位置.与其它DDoS攻击源追踪算法比较,该分步算法具有收敛速度快、路径计算负荷小以及较低的误报率等特点,非常适合实现对DDoS攻击的实时追踪.

基于蜜罐技术的DDoS攻击防御研究

网络信息技术的高速发展,给我们带来很大便利的同时也伴随了各种信息安全问题,其中DDoS就是一种常见的危害严重且难于预防的网络攻击模式,能够短时间内产生洪峰攻击导致服务器因资源耗尽而终止服务。因此已经成为信息安全研究中亟待解决的难题。基于蜜罐技术的DDoS攻击的防御机制,该机制既不需要动员普通电脑用户参与,也不需要ISP服务商支持与配合,只需添加一个蜜罐服务器即可达到比较理想的防御效果。基于蜜罐技术的DDoS攻击的防御技术的原理是通过重定向器重定向黑客对真实服务器的攻击到蜜罐系统中并由蜜罐记录攻击行为数据,为后续制定防御措施提供重要资料。实验中,通过Honeyd开源软件包搭建了虚拟蜜罐系统用于Web服务器DDoS攻击的防御,并对该模型的工作原理做了详细的介绍。