Hybrid classification based intrusion detection and prevention system on cloud environment

Abstract

Cloud Computing systems are widely used for executing computationally intensive parallel applications with different computing needs It provides desirable features and efficient computing facilities to the users through the Internet In cloud computing maintenance of user data becomes the biggest issue due to their sensitive nature The security of cloud computing system is affected by different attacks It degrades the performance of the system in terms of integrity confidentiality and security To overcome these problems Intrusion Detection System is introduced to identify the attacks such as DDoS U2R R2L Probing and Flooding attacks Several types of intrusion detection techniques such as signature based intrusion detection, anomaly based intrusion detection system and hybrid techniques are used to identify the suspicious activity in the cloud computing system. In past, many researchers were concentrated on machine learning approaches for detecting intrusions using artificial network support vector machine, fuzzy clustering and fuzzy with neural network which were not provide efficient result based on detection rate and false alarm rate To mitigate these issues this research work designed hybrid classification based intrusion detection system The proposed hybrid classification based intrusion detection and prevention system is designed with cloudlets, cloudlet controller virtual machine manager and trust authority This work proposes two novel algorithms such as packet scrutinization algorithm and hybrid classifier named NK-RNN The hybrid classifier is a combination of Normalized K-Means Clustering algorithm and Recurrent Neural Network. newline