| Earthquake research involves various kinds of disciplines, such as seismology, geology, and geophysics. It has regional and global characteristics. Therefore realization of seismic information sharing will greatly promote processes of earthquake science. In recent years, techniques of Oracle Database, Application Server and Portal website have been used in seismic information sharing, and a preliminary servier platform for this work has been established in our institute. Although some good results have been achieved, the solution scheme still needs improvement to meet the further requirement of date shar-ing. One possible approach is to use the currently popular technique of Data Warehouse to reach data sharing on a higher level.Since a long time, a great number of seismic and precursor data from stations have been accumulated in our country. How to store, process, and share these data is becoming an urgent problem to be solved. On the other hand, since 1990s the Data Warehouse method has been widely applied to storing and processing of numerous data. Now this technique has become fully mature and got good benefits. Considering the current state of seismic data sharing and the function of-Hata Warehouse, this thesis attempts to make a further study on the seismic information sharing technique on the basis of Oracle Data Warehouse.First, this study analyze, the problems faced by seismic data sharing and their solution schemes, The problems include: (1) Data sources are diversified, complicated, with poor coherence. (2) Amount of system data is exceedingly huge which brings great difficulties to data search and application. And (3) There are various kinds of requirements for seismic data analyses which involve many sections. The Data Warehouse technology can provide a reasonable resolution to these problems. The specificapproach is described below. (1) Make a reasonable data integration strategy by using the ETL(extraction, transformation, load) tool of Oracle to solve the first problem mentioned above. (2) Adopt data partition and suitable index methods to raise efficiency of data management. By predigesting, distilling and uploading useful information from enormously great data to Data Warehouse, the data amount can he reduced greatly and query response is more efficient. This will help solve the second problem. (3) Construct a complete data view in Data Warehouse. On this basis varied Data Marts are set up to meet different needs. This is for the third problem to be solved.Chapter 1 presents the definition, system structure and key techniques of the Seismic Information Sharing Data Warehouse. The key paint of Data Warehouse is to select useful information from many varied systems and perform cleaning to it, so that the data are surely correct. Then through the ETL process (extraction, transformation, and loading) these data are integrated into the data warehouse of one enterprise level, so that a whole view of enterprise data is obtained. Further, using proper tools of query, analysis, Data Mine, and OLAP, these data are analyzed and processed. Finally, the knowledge is present to the managers to provide supports for decision making.Chapter 2 discusses product selection of Data Warehouse. An investigation is made to the popular Data Warehouse products in the current market (Business Objects Oracle, IBM, Sybase, Informix, NCR, Microsoft, SAS, CA),as well as the functions of these tools. The reason for selecting Oracle is then described.Chapter 3 is the important part of this thesis. Taking the construction of data warehouse for underground fluid data, a detailed study on the method and process of establishing data warehouse by usingOracle Warehouse Builder is made. This process includes the following steps: (1) design the structure of data warehouse and data mart, largely for definition of Metadata. (2) crr.ite Metadata, including tables, dimensions, and cubs. (3) define ETL mapping and work flow. (4) finish the deploy and run mapping and work flow. (5) maintain the data warehouse and manage the life cycle by us...
|
PDF Full Text Request