Diverse real-time attack traffic forecasting for cloud platforms

Abstract

Rapidly mounting Distributed Denial of Service bouts is a fatal menace to cloud platforms. Automatic exposure and mitigation techniques are primary defense mechanisms. Identification of attack activities from legitimate network traffic by conventional network traffic monitoring systems are based on statistics. Conventional machine learning techniques are limited with current representational models. In this paper we propose a comparative analysis of hybrid deep learning algorithms and model development for prediction of diverse real-time distributed denial of service attacks. Hybrid deep learning algorithm is a quite effective way of detection and prevention. In this paper, algorithms of machine learning have been evaluated for performance and detection accuracies. We also compare machine learning and deep learning technology used to build forecast on time series ddos data. DDOS data will be provided to produce a forecasting model for prediction of denial of service attacks. This forecast model will be compared with our hybrid model to identify optimal deep learning model for real-time attack traffic detection and mitigation.