Design And Implementation Of Data Warehouse Schedule Archetype System Based On Metadata

The construction of data warehouse contains four processes, including data collection, ETL1, ETL2 and data presentation. There are many complicated relations among these processes and inside them. How to design a better scheduling scheme which not only ensures data correctness but also boosts up the efficiency of system is the question by data warehouse scheduling to solve.The research of metadata and metadata management brings up a new way for the data warehouse scheduling. The CWM proposed by OMG provides a common framework which is used to describe the metadata of data source, data destination, transformation, analysis and management in the field of data warehouse, and offers the standard of metadata communication and share among these isomerous products and tools of data warehouse. Recently, workflow technique is becoming mature increasingly and offers some practical solutions for realizing business process management and process automation and improving the system efficiency.Based on the above background, we study the construction scheme of experimental data center of railroad deeply, by analyzing the practical environment, the present scheduling situation and the problems, we proposed a system of data warehouse scheduling based on metadata. The system which is based on CWM and whose center is the scheduling metadata repository collects all of the data warehouse processes and manages them uniformly. In order to improve the efficiency of system, we use workflow technique to coordinate all kinds scheduling activities. Furthermore we describe the scheduling flow and discuss the key issues of scheme realization, such as representation of operation rule, intercurrent control, task management, task assignment, dependency check and transmit control.