Study And Implementation Of ODS Construction Technology Under Heterogeneous Enviornment

Two types of data processing, operational processing and analytical processing were seperated clearly by data warehousing, so the two-level architecture of DB-DW was come to bring. But the two types of data processing are not entirely different. In actual demands of user, data processing is sometimes multi-levels in which some of operational processings is not suitable for implementation in operational database, meanwhile, some of analytical processings is not fit for implemented based on data warehous. So a new data environment of Operational Data Store was generated, and three-level architecture of DB-ODS-DW was produced.After introducing the related conceptions of ODS, such as definition, characteristics, types, differences and relations among DB, ODS and DW etc., this dissertation mainly studies on construction method of ODS, definition of Meta data, building of Cube, forming of multi-dimensions data report, strategies of data update, implementation of global on line transaction processing (OLTP). An ODS data update strategy based on WAN and heterogeneous environment and a current mechanism of data exchange based on XML are proposed, and then the data consistency of two types environment is guaranteed. While implementing global OLTP, this dissertation proposes two update algorithms to implement the global OLTP applied in ODS and operational database system, and then the real-time update of old data has been realized, and consistency of global data is ensured.ODS is a mixed structure of operational and analytical environment, so it is a more complex system compared with traditional OLTP DB. This dissertation only studies on several key techniques for establishing and managing ODS data integration environment. These key techniques have been applied in practice and perfect effect has been obtained.