基于Hadoop的分布式日志分析系统

陈文波;张秀娟;李林;唐钧

【期刊名称】《广西大学学报（自然科学版）》 【年(卷),期】2011(036)0z1

【摘 要】就如何防御泛洪攻击问题,建立了基于Hadoop的分布式日志分析系统.通过快速提炼访问日志中每个IP地址的请求次数,获取其中请求频率较大的异常IP以有效甄别攻击源,进而为解决泛洪攻击提供重要的现实依据.通过实验,我们不仅验证出分布式日志分析模式较单机模式的巨大时效性优势,还搭建由不同文件系统为基底的虚拟机担当Slaves的两种分布式集群环境,并对两者性能做了详细的对比和分析.%On the issue of defending Flooding-attack, the paper established a distributed system of log analysis based on Hadoop. By the analysis of the number of each IP requests from access logs, we can get the unusual IP that has frequent requests and detect the source of attacks effectively, providing a realistic basis for solving the problem of flood attacks. Through experiments, we not only have verified that the distributed system has a huge advantage of timeliness compared with the single machine system, but also built two distributed systems whose slaves were made up of virtual machines with the base of different file systems, to which performance was compared in detail. 【总页数】4页(P339-342) 【作 者】陈文波;张秀娟;李林;唐钧

【作者单位】兰州大学通信网络中心,甘肃兰州730000;兰州大学通信网络中心,甘肃兰州730000;兰州大学通信网络中心,甘肃兰州730000;兰州大学通信网络中心,甘肃兰州730000 【正文语种】中 文 【中图分类】TP393 【相关文献】

1.基于Hadoop集群的分布式日志分析系统研究 [J], 王润华 2.基于Hadoop集群的分布式日志分析系统研究 [J], 王润华 3.基于Hadoop的分布式日志分析系统 [J], 陈文波;张秀娟;李林;唐钧 4.基于Hadoop的Web日志分析系统设计 [J], 褚龙现;陈婉冰 5.基于Hadoop的流量日志分析系统 [J], 李博;颜靖艺

因版权原因，仅展示原文概要，查看原文内容请购买