A Framework for On-Demand Integration of Enterprise Data Sources

Deploying a data warehouse system in a company is usually an expensive and risky investement. Constructing a data warehouse is a large project that can take very long time and a company cannot know in advance exactly what benefits a data warehouse will offer. Thus, in many cases, data warehousing projects have either been abandoned or been shown to be at least partial failures. We propose a new method by providing a platform to implement business intelligence systems on. The basic idea is to construct the analysis database (i.e. an OLAP cube) on demand and only include the data that is needed for the analysis at hand from the operational databases. In this way the data is always up-to-date, suitable for the current analysis, and some of the biggest risks associated with data warehouse systems can be avoided. In addition, external data can be included in the analysis. The computational costs related to the cube construction are likely to remain at acceptable level, since only the relevant part of the data for the current analysis is needed from operational databases. We outline the use of Grid techologies in the implementation to offer a cost-effective way to harness enough computing power used on parallel processing and sufficient security infrastructure (GSI). To deal with heterogenous data sources the XML language with XSL transformations is applied.