Extending Data Warehouse Dimensions To Reduce Life Cycle Costs
Extending Data Warehouse Dimensions To Reduce Life Cycle Costs (or what companies, and packages…do wrong) It’s widely recognized that a data warehouse is a significant enterprise asset. The goal of the data warehouse is to collect data and make it useful information. ... Summarized Business Intelligence Every data warehouse application is constantly evolving because of changes in business requirements, scope, trends, changes include new management direction and other unanticipated demands of the data warehouse. It may take months to grow an end-user application once the initial data warehouse is constructed and populated into a full decision making environment. ... For a company to exploit an existing investment, the data warehouse design must support continuous change. The existing client application, data, queries, technology, and infrastructure must not change or be disrupted when presented with the demand for new information. A successful data warehouse must cost effectively absorb change. Data Warehouse Basics Dimensional modeling (or data warehouse modeling) is the technique used to design the data warehouse. ... Each dimensions key contributes to the composite key in the fact table. Typical Star Schema Pulling source system data into the data warehouse is probably the most complex single event of the data warehouse project. ... Once the data warehouse is created and loaded, any modifications to ETL (Extraction, Transformation, Load) is very complex and is the biggest contributor to data warehouse life cycle cost. Data Warehouse Infrastructure Two Methods to absorb change Serving as entry points into the data warehouse, dimensions provide access to critical information pertinent to a company’s performance and are key to flexibility and reducing data warehouse life cycle cost. The overall power and flexibility of a data warehouse is proportional to the quality and depth of the dimensions. To absorb change and reduce maintenance cost, dimensions must be capable of defining all potential aggregations (summaries), drill-down navigations and are the basis for constraining and grouping information contained in the data warehouse. However in most cases, as demand for new information evolves, data warehouse dimensions are the first to be impacted and the solution is typically the costly redesign and reprogramming of data warehouse dimensions. Typically, many end-user data warehouse applications fail because the amount of data is too vast for individual analysis. Data warehouse end-user applications fetch thousands of records at a time and should present different levels of aggregation (slices) traversing the dimensional hierarchies.