ETL Research Towards Real Time Data Warehouse

As computer science and information technology developed fleetly, the electronic information data become more and more important in the enterprise daily management. Effective, accurate and timely Data analysis is required imminently by enterprises. Traditional data warehouse commonly just support some analysis and queries on history data while the commerce information variety can't be displayed in real-time. Real-time data warehouse extends the application of traditional data warehouse and can support tactical queries for enterprise. Real-time data warehouse that effectively cut the time-lapse can provide much variable tactical decision support. ETL is the most important part in Data Warehouse constructing, it's become more complex and tight to implement real time ETL in Real Time Data Warehouse. Therefore, the investigation towards real time ETL solution is very weigh for the Real Time Data Warehouse implementation.The paper at first introduces the background of the study on technology of Real-time data warehouse ETL, actuality of real-time data warehouse system, Through contrast between the real-time data warehouse and traditional data warehouse, a series of countermeasures to realize the real-time data warehouse ETL were given. Then according to the real-time data warehouse structure, this paper discussed the central issues about changed data capture and aggregate data update, and advanced new solutions as adopting data replication and real-time segment. Synchronously, the design and implementation of the two solutions was discussed in detail. Lastly, this paper introduces the design and implementation of an experiment based on Book publishing system. The solutions are proved to be reasonable and effective by the experiment result.