Towards MapReduce based classification approaches for Intrusion Detection

Abstract

The term Big Data is explosion of high frequency digital data encountering daily through various sources. Velocity, Volume, Variety, Veracity and Value is causing difficulty for processing, storing and analyzing the Data. Intrusion Detection System in Big Data environment is one of the research issue we addressed. Intrusion Detection is a security technique, used to monitor and analyze network traffic in order to detect network violation. We require a robust Intrusion Detection technique to classify between normal and anomalous data and predict security breaches. In this paper, we have analyzed Machine learning techniques to detect intrusion which can scale up to build such systems. There are many algorithms one can opt for depending upon the need of system. This paper deals with Naïve Bayes and K-Nearest Neighbor classifier in MapReduce framework and their performance comparison with WEKA implementations. Our preliminary analysis over NSL-KDD seems to be promising.