中文摘要：

分布式拒绝服务（DDo S）攻击是一种近年来在网络上流行的能够导致巨大损失与破坏的拒绝服务攻击,攻击方式正以多样化智能化快速发展.因此,DDo S攻击的检测技术越来越成为研究的重点,如何准确及时的识别攻击成为检测技术要迫切解决的问题.近几年,很多人或组织致力于DDo S检测的各种关键技术的研究中,也取得了显著成果,但这些算法有些地方仍需要改进,本文以实时检测DDo S作为目标,深入分析DDo S的特征,对DDo S攻击的检测技术进行了研究和现实分析,采取了基于信息熵的协同检测算法.在局部检测中对目的 IP地址和源IP地址进行统计分析,采取子空间与K-means算法相结合的方式估算信息熵,然后对检测信息及信息熵进行融合,采取全局决策的方式来达到检测的目的,通过实验进一步验证了论文中所提出算法的优越性.

英文摘要：

Distributed denial of service （ DDoS ） is one of the denial of service attacks which are popular and leading to great economic loss and damage. The ways of attack are more diversified and intellectual. Therefore, DDoS attack detection technology has increasing- ly become a focus of the study, and how accurate and timely detection techniques to identify attacks become pressing problems. In re- cent years, manys people and organizations are engaged in research of all kins of key technologies of DDoS detection and hava got out- standing achievements. But the DDoS detection still need to be improved. With the aim of real-time detection of DDoS, the article makes a detailed analysis of characteristics of DDoS, research about detection technologies of DDoS and adapts collaborative detection based on information entropy. Do statistic analysis of destination ip address and source ip address in the partial inspection, by then method of subspace and K-means to estimate information entropy and then integrate the detection information and information entro- py ,to detection in the global decision way to further prove the superiority of the algorithm.