IMPLEMENTATION OF SDN BASED FEATURE SELECTION APPROACHES ON NSL-KDD DATASET FOR ANOMALY DETECTION

In past decade, traditional network was used for transferring of data between nodes. The main issue related to traditional networks was their stable nature and also, they were unable to meet the requirements of newly added devices in the network. So, traditional networks are replaced by Software Defined Network (SDN). Many of the networking applications rely on network for transfer of data. SDN networks are dynamic in nature. SDN can be used to create a framework for data-intensive applications like big data etc. Now a day, security of data over the network is very crucial. Machine learning (ML) algorithms are used for classification of network data in order to detect intrusion attacks. In this paper, a comparative analysis of machine learning algorithms is done by using different feature selection approaches. For this analysis, NSL-KDD dataset from training and testing with 41 features and 125000 samples are used. Accuracy estimation of machine learning algorithm with a particular feature selection approach can is done in order to detect anomaly over SDN.