A Framework Study of ETL Processes Optimization Based on Metadata Repository

摘要

ETL plays a significant role in DW construction. Basically,ETL management is rooted in metadata management of the ETL processes. A metadata management system with good design can highly improve the ETL efficiency. However,the large amount and the wide distribution of metadata in ETL processes cause the mismanagement of metadata,which hasn't got a good solution yet. To deal with this problem,the paper suggests intensively manage ETL by metadata repository. Metadata repository can show DBA the metadata in a direct,simple,and centred way,and makes metadata more easy to understand,therefore metadata management becomes more direct,simple,and centred. Compared with the common one,ETL processes based on metadata repository have a much better optimization effect.