Analysis of Big Data for Detection of Network Intrusion

One of the main challenges associated with analysis of big data is automatic detection systems that classify network traffic data. The aim of this paper is to consider design and implementation of intrusion detection systems (IDS) using several classification algorithms for big data analysis. Big data analysis techniques can extract information from a variety of sources to detect future unknown attacks. We use classification algorithms with MapReduce framework for mining IDS in Apache HTTP server on a Linux system. So that, K-Nearest Neighbor (KNN), Support Vector Machine (SVM) and Linear Discriminant Analyses (LDA) classifiers are implied on NSL-KDD Dataset and compared them with some wellknown existing techniques for IDS. The results show that the average efficiency is high. The Minimum efficiency reportingvalue is 95% and maximum 97% by changing the parameters in the proposed model.

Keywords:

Big data , Detection , HDFS , MapReduce , NSL-KDD , KNN , SVM , LDA.