基于大数据分析的DDOS检测系统的设计与实现  

Design and Implementation of DDOS detection system based on big data analysis

作　　者：金磊[1] JIN Lei(Xinjiang Aksu Institute of Education,Xinjiang Aksu,84300)

机构地区：[1]新疆阿克苏教育学院

出　　处：《自动化与仪器仪表》2018年第11期121-124,共4页Automation & Instrumentation

摘　　要：随着网络应用服务量的增长,网络安全事件也呈现爆发式的增加。为了对大数据下网络DDoS攻击进行检测,文中以Spark、Flume和HDFS为基本框架,设计了一种大数据分析的DDoS检测系统。采用了apache spark分布式计算框架,结合数据采集层、存储层、计算层实现了对DDoS检测系统的框架设计。并利用Bro可以通过扩展的结构化日志对网络活动进行记录,从而实现对网络漏洞的检测。同时,采用Corsaro的网络分析器实现对目标IP的分析与记录。最终,在Spark实现了基于贝叶斯分类器和cart决策树的DDoS的检测,并对其进行了验证。With the growth of network application services, network security incidents also show explosive growth. In order to detect network DDoS attacks under large data, this paper designs a DDoS detection system for large data a- nalysis with Spark, Flume and HDFS as the basic framework. This paper uses the Apache spark distributed computing framework ,combines data acquisition layer, storage layer and computing layer to design the framework of DDoS detection system, and uses Bro to record network activity through the expanded structured log, thus realizing the detection of network vulnerabilities, and using Corsaro network analyzer to realize the target. The analysis and record of standard IP. In this paper,we have implemented DDoS detection based on Bias classifier and cart decision tree in Spark and verified it.

关 键 词：大数据 DDOS攻击 网络安全 网络漏洞 贝叶斯分类器 

分 类 号：TP393.08[自动化与计算机技术—计算机应用技术]