Detecting DoS Outbreaks in Cloud Environment Using Machine Learning Algorithms in Hadoop Cluster

The necessity to handle data has grown critical since the introduction of cloud computing and continuous Internet access. In addition, the number of network intrusions has grown. As a result, developing a defence system for cloud networks capable of detecting breaches is critical. To build a successful intrusion detection system (IDS), the features selection approach is essential. To classify the network traffic in less time and with enhanced accuracy, metaheuristic algorithms were employed to reduce the number of features. The particle swarm optimization (PSO) and whale optimization algorithm (WOA) are combined in the proposed method for detecting denial-of-service (DoS) attacks to solve the feature selection (FS) concerns. WOA with high exploration capability avoids PSO to fall into local optimum. This research shows an innovative machine learning method implementation for DoS detection in Hadoop-based clusters. The random forest (RF) methodology is used to classify data. The outcomes have demonstrated that the proposed approach successfully improved the PSO algorithm while employing the WOA. The features selected by PSO-WOA from UNSW-NB15 dataset, and classifying them using RF attains maximum accuracy of 93.88%, sensitivity of 91.70 and F1-score of 95.44. The studies reveal that the proposed system outperforms the system based on state-of-the-art approaches.