A Research On Intrusion Detection Algorithm Based On Software Defined Security Service In Cloud Environment

In recent years, cloud computing technology has been developing rapidly. But cloud computing security problems have hampered the further promotion and application of cloud computing. Therefore, many researchers, organizations and cloud security vendors have devoted to the research of cloud security, put forward many solutions. With the rise and popularization of software defined network technology, this paper proposes a cloud security solutions based on software defined network, analyzes the operation mechanism of the solution, and the solution is different from the other solution with software defined, on-demand, focus and panoramic four characteristics, the security as a service is no longer a fantasy, but the landing feasibility, namely in the scheme of security as a service, it will calculate the configuration and deployment in the cloud environment according to the needs of users, in order to achieve a safe "software defined".Then, the paper focuses on how to implement intrusion detection method-- clustering analysis technology. In the study of cluster analysis, we focus on the hard C-means algorithm and fuzzy C-means algorithm, then according to the existing algorithm in intrusion detection of defects, and put forward a combination of generalized regression neural network algorithm which called FCM-GRNN algorithm, and based on the algorithm design of the intrusion detection system based on software defined security. GRNN neural network has a faster learning rate and good convergence and the learning curve smoothing is not easy to shocks, especially when dealing with large amount of data samples, using GRNN can achieve fast approximation and is very effective in the real-time processing sparse data. Therefore, the combination of GRNN and FCM can not only give full play to the advantages of both, but also meet the needs of the cloud computing environment need to deal with massive amounts of data. In order to verify the feasibility of the algorithm, efficiency and superiority, in this paper the final design of the two sets of experiments, through the simulation experiment fully comparing the detection of HCM algorithm, FCM algorithm and FCM-GRNN algorithm in intrusion detection rate, false alarm rate and large amount of data processing in time, it is concluded that the FCM-GRNN algorithm for the detection rate is higher, the false positive rate of the lowest conclusion, and and the other two algorithms compared to algorithm FCM-GRNN more applicable to intrusion detection in the cloud environment.