Created a Data Warehouse with Kimball methodology, conformed dimensions using best practices recommended by Informatica and Oracle. About half a terabyte of historical information and increasing every day. By shrinking the pipeline at every stage of data load and with partitioning, parallelism and de-coupling the user tables from data warehouse, we are able to achieve almost 24x7 availability with the exception of cold backups on weekends.
Now that there are pre-built data warehouse data models and tools that are more advanced and can scan through huge amounts of data (big data), data preparation and data warehouse could be split separately and data can be kept more current, as opposed to one load a night.
