| With the coming of big data, many organizations are collecting data from a variety of sources to analyse and make decisions based on the analysis. The processing platform Apache Hadoop is becoming more and more popular. However, while widely used, Hadoop is gradually exposed many security issues. Therefore, from 2009, Apache project devoted a team to undertake an enhanced security mechanisms work on Hadoop to increase security authentication and authorization mechanisms based on ACL.But it is often complex to deploy Hadoop platform. The Manager for Hadoop system is developing rapidly. It realized to help to create clusters, monitor clusters and configure clusters so that it can make it easier for administrators. And the building of the subproject Authorization System is also very important for comprehensive management of security.If companies only use Hadoop open source cloud computing framework, the cluster administrators need to set access security properties in each Hadoop component one by one. And the tasks are all done manually. It is so inefficient and lacks flexibility.Therefore, we developed the Authorization System based on RBAC(Role Based Access Control) model. It adopts some advanced technologies such as GWT, GWT RPC and JAXB. With the role-based access control, users are no longer directly linked to permissons. And they are assigned to roles, which are associated with permissons and can be created or deleted by the administrator. Using JAXB, we can save the authorization information to xml file, and achieve mutual conversion between Java objects and xml files. It is very convenient and flexible. With GWT, we can replace JavaScript with Java programming language for front-end interface.Administrators can manage the authorized users of the system, can enable or disable the access control mechanism for each service, can configure role properties by the authorization system. And the system can process authorization configuration and deploy it to Hadoop cluster automatically. |
PDF Full Text Request