A Data Warehouse provides the supporting infrastructure for successful Business Intelligence initiatives. However, exactly what constitutes a data warehouse and the reasons why you should build one can be confusing.

Similarly, the difference between data warehouses and data marts is not always clear.

Do I need both? Why?

There are several different viewpoints, most of which are variants of two different architecture concepts.

The Corporate Information Factory is probably the most well-developed concept that stresses the importance of building both a data warehouse AND data marts. This concept, developed by Bill Inmon and Dr. Claudia Imhoff is well regarded and has proven to be very successful.

Equally successful and well respected is the theory, developed by Ralph Kimball that a data warehouse is simply a collection of integrated data marts. On the face of it, these are two very different viewpoints, but when you analyze them in depth, there are some striking similarities. Kimball includes a ‘staging area’, where the source data is staged before loading into the data mart tables. In effect, this is equivalent to the data warehouse in the Corporate Information Factory model.

While there are other differences in concept, terminology, methodology etc, these two dominant architectures both agree: to be successful in Business Intelligence, you MUST implement a well architected, conformed data repository and use ETL (Extract, Transform, Load) processes to load it from your source data.

This is EXACTLY what RODIN is designed to do. 

Whether you are a fan of Ralph’s, or of Bill and Claudia: if you want to build a data warehouse and/or data marts on your System i server, RODIN is the solution you are looking for.

Building and managing tables and indexes normally requires skilled programmers and DBAs. With RODIN you are removed from the complexities of DDS or DDL, and can concentrate on the business entities and information that you need to store. Similarly, ETL typically involves complex programming, with a whole raft of associated issues such as error management, auditing, change control, documentation, metadata – all of which combine to make this a huge challenge.

Sadly, these complexities are usually very much underestimated and many organizations do not implement an ETL tool. This almost always results in project overruns, or more commonly reduced function. Metadata, audit reporting, error management and documentation are typical casualties.

RODIN takes away this complexity and provides a complete development and deployment environment, with fully integrated processes and metadata. Using RODIN, development time is reduced significantly, while at the same time, delivering all the error, auditing and metadata reports you could wish for (and more).