摘要
随着计算机网络的应用和普及,人们对网络安全的要求也愈发强烈。在网络安全领域中,DDoS被认为是最具有破坏力的网络攻击之一,它的破坏力随着网络带宽的增加而增大。针对于传统的网络层DDoS攻击检测和预防手段已趋向成熟,攻击者则向更高层次——应用层DDoS攻击转移。应用了无监督学习方法对应用层DDoS攻击进行检测,并对具体的网络场景进行数字仿真,通过对仿真结果的分析,判断出被攻击的网络区域。通过对不同类型访问者在网络中神经元分布情况的分析,着重指出攻击者与正常访问用户的访问行为间的区别,为日后进一步对应用层DDoS有效检测和防范提供理论基础。
With the application and popularization of the compu%er network, people have increasingly intense for network security requires.In the field of network security, DDoS attacks are considered one of the most destructive of the network attack, the destructive is more powerful with increasing of the network bandwidth. For the traditional network layer DDoS attack detection and prevention tools are maturing, the attacker then to a higher level - application-layer DDoS attacks. In this paper, a method for unsupervised learning to detect application-layer DDoS attacks, and specific network scenarios for digital simulation, through the analysis of simulation results, determine the areas of the network being attacked. Through the analysis of different types of visitors in the distribution networks of neurons, highlighting the difference between the attacker access and the user's access behavior. Through these provides a theoretical for the DDoS application layer detection and prevention in the future.
出处
《计算机安全》
2014年第6期11-14,共4页
Network & Computer Security