Research On Key Technology For Sharing Distributed Data On Meteorology

It is quite important to build a system for sharing and management of the distributed meteorological scientific data. To establish the system, there are two highlighting issues to be resolved: the access control mechanism for the distributed data and the implementation of the directory service. As the core module of the security model, the access control mechanism is the guarantee of the data safety; And as the core function of the sharing and management operation, the directory service is the guarantee of system work. The system can run safely and sequentially only by solving the two problems above in the same time.This paper concerns about two aspects of the security model: how to make sure the data can be accessed only by the authorized users and how to minimize the complexity of the user access under the premise of the system security. After the study of the RBAC model, we find that it may bring role conflicting associations and enhance the permissions of the role covertly. Therefore, this paper presents a model of RBAC which based on trust-domain. We can solve the problems above effectively by applying the mechanisms of roles-recommended and unidirectional role-mapping.Metadata management system is the core function module of the meteorological scientific data sharing and management system, and directory service is the core service of the metadata management system. This paper analyzes the directory service model based on LDAP, and proposed a new LDAP model based on metadata. In this model, the resource information for the dataset kept in the meteorological departments is separated with the registered metadata information, and each node has a full control of its metadata directory. Therefore, user node can modify the resource information, and the metadata directory in the resource centre doesn't appear the migration problem. Calculating the Hash value according to the keywords reduces the query complexity and has better scalability and availability.According to the above research and the characteristics of meteorological scientific data, we design and implement a system for the sharing and management of distributed meteorological scientific data based on the SIMDAT open source project. The scalable and configurable system developed with several modules lays the foundation for the sharing of the meteorological scientific data.