Metadata And Its Management Of DM OLAP Server

Metadata has been identified as a key success factor for Online Analytical Processing(OLAP) server. It is the most important driver to increase the efficiency of the use, development, maintenance and management of an OLAP server. The project studies several issues on the metadata management of the OLAP Server DM_OLAP, including the generation, storage, update, maintenance and share of the metadata.According to the requirement of user's multidimensional data analysis, we design an private object-oriented multidimensional metadata model. It organizes data cube definition into DM_OLAP metadata. This model also separates metadata applications from data dictionary, with the result that metadata application is simplified. A relational schema for the implementation of this metamodel is proposed to store the metadata in DM3 relational DBMS. Based upon all these works, we develop the DM_OLAP data cube defining guide.The update and maintenance of metadata relate to several key problems of metadata management in data warehouse, which include access control, update, version management and update propagation. Having analyzed the organization and storage structure of the metadata in DM_OLAP, we use the three technologies of role based access control, version control in engineering database and lazy two-step data synchronization to solve these problems.Being a key component of the Decision Support System(DSS), it is very important for an OLAP server to share its metadata with other components in the distributed DSS environment. We choose the federated metadata share approach to solve this problem. So we analyze the Common Warehouse Metamodel(CWM) to find the correspondence between the CWM OLAP package and the metamodel of DM_OLAP. Based on this correspondence and the DTD defined according to the XML DTD production rules of XMI, we design several algorithms to develop the DM_OLAP metadata export tool, which can transfer the metadata into XML files.