| In medium-term of 1980s, the father of data warehouse-Bill Inmon made a definition of data warehouse in his book of"Building The Data Warehouse". Once the concept of data warehouse was made, it was applied to the main traditional data-intensive industries,such as telecom, bank and insurance. In foreign countries many large data warehouses were established in 1996-1997. With the building of data warehouse, DB-DW structure was gradually shaped in enterprises. However, with the appearance of requirements on multi-level data processing, the DB-DW structure was incapable of satisfying with the enterprises' requirements. Then Bill Inmon defined the concept of ODS (Operational Data Store) in his following book of"Building The Operational Data Store",which was used to meet the daily demand of analysis, decision and management for enterprises. The introduction of ODS took shape of DB-ODS-DW structure.Under the pressure of industry competition and business requirements of China Insurance Regulatory Commission, China Life Insurance Corporation intends to establish a Statistical Information System with data warehouse technology. In view of the business requirements and actual applications of China Insurance Regulatory Commission and China Life Insurance Corporation, in this thesis we discuss the problems and solutions of designing the architecture of Statistical Information System. Furthermore, we design the architecture on the basis of DB-ODS-DW structure, and research on incremental update strategies on ODS, and discuss how to update data incrementally according to different kinds of data in ODS in the case of massive data, and especially, apply DROV (Differential Re-Evaluation Of Views) algorithm to update materialized views incrementally. For the project of statistical information system of China Life Insurance Corporation, the DROV algorithm is implemented on the province's ODS system and high efficiency is achieved. |
PDF Full Text Request