I must be using the wrong term. I want to describe a database where I will store all my Corporate Data (used in multiple systems) for use -- not on a reporting basis, but as source data to other systems.
To me this is a Data Store (which I have been mistakenly calling an operational data store).
What would the term be, if there is one, to describe this?
There are so many definitions and terms floating around in business intelligence (BI) that it is impossible to be definitive. Please regard anything I write as "more of what you'd call "guidelines" than actual rules," as Barbossa says in "Pirates of the Caribbean: The Curse of the Black Pearl."
Having protected my back as much as possible...
An operational data store (ODS) is a place where data from multiple source systems is stored. As a general rule, an ODS is maintained in near real time and the data is usually at the transaction level.
Assuming that you don't intend to do the 'real time' thing then you are probably wise, as you suggest, not to call it an ODS.
So, if it isn't an ODS, what is it? Well, I would argue that it depends on the purpose you have in mind.
If you intend to use it for direct reporting, then the best term is probably enterprise data warehouse (EDW).
If you are going to feed data from it to data marts for reporting, then the term "data warehouse" is probably best.
However, you say "I want to describe a database where I will store all my Corporate Data (used in multiple systems) for use, not on a reporting basis, but as source data to other systems."
Well, that's OK. If the feed from it to the other systems isn't real time but batch, then the term "data warehouse" is probably still OK because I know of several data warehouses that are also used as data sources in a limited way.
However, you also say "My end thought is I will eventually be able to minimize any system's need for interfaces to the one that will connect to this data store system."
That sort of implies that you intend to use the data store you are creating as an operational database. That would be a huge undertaking since the new data store would have to mimic the business logic and functionality of all the original systems. So I have to admit that I don't know of any general term that would apply; mainly because so few people would attempt this.
This was first published in April 2008