Difference between NDS and ODS.

Hi,

I am a little unclear as to the difference between a normalised data store and an operational data store. I thought that the main difference was that a NDS was an internal data store (not seen by users/applications), and an ODS was a hybrid data store, used as both an internal store before being moved to a dimensional data store, but also one that could be seen by users and applications. I have now read that an ODS does not contain any historical data, and contains only current transaction and master data. Is this right?

Regards,

M.

Unfortunately, there aren't any standard definitions for these labels. In fact, I haven't ever used the term normalized data store. About all you can reliably say about and ODS is that is contains operational data that may be used for near real time reporting. How it's modeled and how much history is contained in it are subject to personal implementation. I personally like several weeks of history in an ODS to facilitate reloading the warehouse in case I find a bug.

Agree. NDS doesn't mean anything other than a normalized database. Inmon uses the term Enterprise Data Warehouse to describe the primary data store in his architecture. He is a proponent of a normalized EDW with publication to data marts (dimensional or otherwise) for end-user consumption.