Research On Intrusion Detection Method For Mobile Intelligent Terminal In Cloud Environment

With the rapid growth of the internet technologies such as mobile intelligent terminals and cloud computing over the last few years,the integration of those two technologies has given rise to the new concept called Mobile Cloud Computing(MCC)and it has got a wide attraction in both academia and industry.According to Cisco IBSG(Online 2016),close to 85% of the world's population has access to mobile terminals.However,MCC due to its distributed nature,wide user access and easy to use enable intruders to exploit cloud services for their advantages without the permission of the system administrator.Meanwhile,more security methods such as firewall technology and intrusion detection systems(IDSs)have been used by many researchers to defend the cloud computing problems,but still have some limitations of lack of scalability and self-adaptability for firewall technology,and low detection accuracy,low true positive rate,high false positive rate and redundancy issues for IDSs.Aiming at the aforementioned deficiencies,this thesis mainly uses some different machine learning approaches-based IDS such as classification and information theory to deal with intrusion detection problems and overcome the drawbacks of firewall technology and other traditional intrusion detection methods.The classification-based IDS and feature selection methods adopted in this thesis include support vector machine(SVM),random forest(RF),information gain(IG)and MapReduce for Evolutionary feature selection(MR-EFS).In this thesis,the main research works are as follows:1.To improve the classification accuracy,detection rate and reduce the high false positive rate,the support vector machine(SVM)and random forest(RF)classifiers were used for binary classification in either normal or malicious attack.2.To deal with the redundancy issues,the information gain(IG)based feature selection and MapReduce for Evolutionary feature selection(MR-EFS)were used.The mostly used datasets to analyze and evaluate IDSs problems like KDD'99 and NLS-KDD datasets,have been facing a serious problem of large amount of unimportant data including redundant and irrelevant data,in this thesis,the IG based feature selection and MR-EFS have been used to remove the redundant and irrelevant features,and reduce the computation complexity of these datasets.3.Intrusion detection system as a security management tool has been adopted by many researchers over the last few years to control different events occurring in a computing system or network and analyzing them to find the malicious activities either within the system or outside the system,and in this thesis,the IDS methods have been proposed to control and analyze those events in mobile cloud computing(MCC)and through the experimental results,our proposed methods have detected the intrusions in MCC with high accuracy,detection rate and low false positive rate.