Font Size: a A A

Research On Hadoop Security Access Control Mechanism Based On User Behavior Assessment

Posted on:2018-04-02Degree:MasterType:Thesis
Country:ChinaCandidate:L X MengFull Text:PDF
GTID:2348330533960101Subject:Computer technology
Abstract/Summary:PDF Full Text Request
As an open source distributed computing cloud platform,Hadoop has been widely applied by the E-Commerce and Internet companies with its unique characteristics of high reliability,high scalability,high efficiency and high fault tolerance,etc.At the same time,its security issues are also increasingly prominent.In numerous cloud security issues,data security is one of the core issues.Access control technology has become an important guarantee for data security in cloud platform by restricting users' ability to access data and thus ensuring that resources are not illegally used and accessed,at present,the research on security access control mechanism of cloud platform has become a hot spot.Based on the research of the security mechanism of existing Hadoop platform,a dynamic access control model based on user behavior assessment(DACUBA)was proposed combined with existing Hadoop platform security mechanism to solve the problems that existing Hadoop platform access control mechanism design does not consider the user's normal or abnormal attribute changes,lack of dynamic and other defects,.this model dynamically adjusts the user's access rights and improves the security of cloud platform through real-time monitoring and assessment of user behavior of cloud platform.Firstly,Hadoop existing security mechanism was analyzed,and DACUBA model was proposed,the access control elements,rules and procedures of DACUBA model were designed based on the Hadoop existing access control mechanism,the concept of user behavior assessment was introduced.Then,according to user behavior assessment rules,This paper proposed a method of user behavior evaluation,which uses the parallel command sequence learning(PCSL)algorithm to extract the user instruction behavior contours,establishes the global K model model library to classify the sequence of user command behavior,and designs the assessment formula to calculate the user comprehensive assessment value using classification result.Finally,based on the above theory and Hadoop source code,DACUBA model prototype system was designed and implemented.The Hadoop experimental platform was set up in the laboratory environment.The behavioral classification,behavior assessment,feasibility,effectiveness and performance costof the DACUBA model were verified using the modified Schonlau data set.The experimental results show that the behavior classification method of the model is more accurate,the evaluation method is reasonable,and the time,performance cost is small.The model access control strategy can effectively implement the dynamic access control to the cloud platform users in combination with the behavior assessment algorithm.
Keywords/Search Tags:Cloud platform, Kerberos, Hadoop, User behavior, Access control, Parallel command sequence learning, Security
PDF Full Text Request
Related items