Font Size: a A A

Research On Dynamic State Detection And Static State Evaluation Technology Of Cloud Computing Security

Posted on:2018-06-30Degree:DoctorType:Dissertation
Country:ChinaCandidate:X L ZhaoFull Text:PDF
GTID:1318330542479151Subject:Computer application technology
Abstract/Summary:PDF Full Text Request
With the wide application of cloud computing technology,more and more attention is focused on the security of cloud computing.How to determine the security between several cloud computing solutions has become the problem of cloud computing users are bound to face;How to dynamically detect the security of cloud computing in the process of cloud computing has become a hot topic in the research field.In this paper,the cloud computing security faces of these two aspects merge to from the dimension of time to look at security evaluation,the evaluation scores for the static evaluation and dynamic time face detection for.Static evaluation can solve the pros and cons of cloud computing methods to determine the means of dynamic testing to become cloud consumers understand the security of the operation of the process of cloud computing tools.In the dynamic detection,there is no better way to detect both known and unknown intrusion detection methods.In the static evaluation,there is noquantitative evaluation method for cloud computing security.In this paper,a hybrid intrusion detection method and quantitative evaluation method for cloud computing security is studied in two aspects.For the cloud computing of hybrid intrusion detection problem,based on evolutionary algorithm,an improved unsupervised learning k-means method and the application silhouette coefficient and subspace clustering method,according to the unsupervised learning feature to construct detection mode,and puts forward to establish cloud computing distributed intrusion detection model,a comprehensive cloud computing hybrid intrusion detection method,EASKS.At the same time the simulation experiment and practical project deployment,the method is better than that of neural network algorithm,particle swarm optimization algorithm and the standard k-means algorithm,and can detect unknown attack and can improve the known attack detection rate and detection speed.It is proved that the cloud computing dynamic detection safety effective.Cloud computing service providers and cloud consumers can improve the ability of cloud computing intrusion detection through the application of this method.There is no generally accepted security index system of cloud computing,to based on an index system to conduct a comprehensive study,to traditional information security level protection evaluation based,according to the characteristics of cloud computing system,analysis obtained CIAAAT model through to threat,vulnerability and risk modeling,formalization of security risk calculation method,corresponding to the index system constitute the basis of safe control point,based on the established characteristics of safety evaluation index system for cloud computing,and applied in the actual project,with traditional methods do compared to prove that the system is for the cloud computing is more applicable.For cloud computing at a certain point of time safety evaluation work in,non uniform dimension index,index coverage,indexes exist some mutually exclusive,combined with grey relational analysis method and VIKOR method to form a quantitative evaluation method GRAVIKOR.And based on cloud computing security evaluation index system,the use of GRAVIKOR method to form a cloud computing security quantitative evaluation value.Finally according to the actual case of quantitative evaluation and analysis,using the method in safety evaluation in cloud computing,more practical than the traditional method,also can take into account the preference of experts,but also can guarantee the objectivity of the evaluation results,so as to solve the problem of cloud computing static safety evaluation problem.Based on cloud computing data distribution system,the application method of dynamic detection and static evaluation in real environment is studied.Through the application of traditional static evaluation method and GRAVIKOR method in this system,the results of the two methods are compared and analyzed,and the security short board of this system is found;By deploying an intrusion detection system using the EASKS method,the system is attacked for a period of time,and the data is analyzed from a dynamic perspective of the system security short board.Based on the comprehensive analysis of the security short board obtained by static evaluation and dynamic detection,the security short board of the system is obtained from the global perspective,and the improved method is given for the corresponding short board.Through the actual application research of cloud computing system,the effectiveness of EASKS method and GRAVIKOR method is verified,and the scientific analysis basis and improvement method are provided for users through the application of the method.Through the research on cloud computing security of static evaluation and dynamic detection method proposed cloud computing security quantitative evaluation method and fill the cloud computing security research blank,and can become discriminating cloud computing schemes,which provides an effective tool for the third-party evaluation institutions and cloud consumers.Cloud computing security evaluation index system is proposed,which provides a basic method for establishing evaluation index system,and lays a solid foundation for advancing the security evaluation of cloud computing.Cloud computing oriented unsupervised learning algorithm for hybrid intrusion detection method based on solve the cloud computing oriented virtualization,visits,large data volume rapid discovery of known and unknown intrusion is put forward in this paper.The method can autonomous learning and constantly improve the detection rate,which provide a means to detect cloud computing operating state security to cloud consumers and cloud operators.
Keywords/Search Tags:Cloud Computing, Intrusion Detection, Quantitative Evaluation, Semi-supervised Learning, Multiple Attribute Decision Making
PDF Full Text Request
Related items