Whenever you create a data warehouse for reporting, you have to consider that there are finite resources. There is never enough space to store data, never enough time to populate the database, and never enough processing power to use a fully normalized source. Even with the latest superfast technologies, such as Oracle Exadata, there is a limit to the amount of data that can be stored or processed in a given time period (ETL window).
The primary objective should always be to speed up the report production, which means using a dimensional model, particularly when storing a large amount of data. As discussed in the Theories and models section, using an entity relational database would reduce the redundancy of data and, therefore, reduce the amount of data stored and the time taken to load it. However, the report production time would increase.
It is always necessary to establish a trade-off between data volume (which improves speed of reporting...