Research On Hadoop Platform Dynamic Access Control Model Based On Risk

With the arrival of the big data era,Hadoop provides a platform for distributed storage and high performance computing.As more and more data are stored and calculated on Hadoop,the security of Hadoop has received extensive attention.Access control is one of the important technology for ensuring data security.By accessing authorization,Hadoop security can be effectively improved.The risk-based access control model is characterized by strong dynamics and high flexibility and is widely used in cloud platform.Therefore,the Hadoop platform dynamic access control model based on risk has become a hot research topic.Firstly,the existing Hadoop access control model was summarized and the existing cloud platform risk-based access control model was analyzed.The programming regression model,event calculus(EC)mechanism,and the Kerberos of Hadoop access control mechanism were summarized and analyzed as a good foundation for the following research.Secondly,aiming at the problem that the weight of risk assessment index in cloud platform risk access control model was static and subjective,an adaptive risk assessment index weight distribution method was proposed,the risk assessment index weight adaptive distribution model was designed,the programming regression algorithm was proposed to solve the weight,and a risk value quantification formula with adaptive weight distribution was constructed.On this basis,a dynamic risk access control model with adaptive weight distribution was proposed.The experimental results show that in terms of accuracy and sensitivity,the calculation of the risk value is better when the model is dealing with large concurrent users and the response time is shorter,which is more suitable for cloud environment.Finally,as the risk access control model couldn't match rules dynamically and the risk values were insensitive to access requests,a cloud platform dynamic risk access control model(CDRAC)was proposed.The attribute-based access control(ABCA)was improved with EC mechanism,and the dynamic rule matching module was constructed in CDRAC.According to an adaptive risk assessment index weight distribution method,the risk assessment module with high sensitivity to access requests was constructed in CDRAC.Experimental results show that CDRAC has good effectiveness and feasibility,and it hasbetter real-time performance and adaptability than other existing models.