The Comparative Analysis between the Dimensional Modeling and the Entity-Relation Modeling
Ion Lungu, Mihaela Muntean · SSRN Electronic Journal · 2006
On-line transaction processing systems (OLTP) is profoundly different from dimensional data/warehouse, the basis for open decision support systems. Both OLTP and datawarehouse systems are greatly concerned with data consistency However OLTP consistency is microscopic. The point of transaction processing is to proces a very large number of atomic transaction without losing any of them. A OLTP system processes thounsands or even millions of transactions per day. Each transaction contains one small piece of data. In a datawarehouse, consistency is measured globaly and often will process only one transaction per day. We don't care about an individual transaction but we care enormously that the current load of new data is a full and consistent set of data. We have a quality assurance perspective and a new job is created the data quality assurance manager. This transaction contains thounsands or even millions of records. We call it the production data load. Dimensional modeling (DM) is the name of a logical design technique often used for datawarehouses. It is different from entity -relation modeling (ER) used for OLTP systems. This article points out the many differences between the two techniques.