Framework to Analyze Malicious Behaviour in Cloud Environment using Machine Learning Techniques
Pranay Jha, Ashok Kumar Sharma · 2021
The era of digitalization adds a lot to indorsing an IT infrastructure to meet business needs. Cloud Infrastructure played an important role by offloading the on-premises workloads to the datacenter which is easily accessible from anywhere through internet and reduce the complexity. It poses a major security concern while storing the data over internet. There are already a lot of security frameworks to protect the cloud environment, notwithstanding, there has been a lot of risk in recent years that has increased security threats and breached data at even higher levels. In this paper, we have proposed a framework to provide the security of cloud environment based on user's Behaviour. It counsels an improved method for detecting normal and malicious Behaviour and prevent the attempt which is coming as abnormal access. We used UNSW-NB15 dataset which contains more than 82,000 samples, 49 features, and 9 types of attacks. We have presented the viable analysis of performance of different machine learning techniques used for prediction of user's Behaviour and comparison of various algorithms i.e. Support Vector Method, Decision Tree, Naïve Bayes, Random Forest, KNN and Logistic regression, which shows the promising results in detecting the malicious behaviors.