Reducing duplicate streams and databases

kduke · Post by **kduke** » Tue Mar 27, 2007 5:57 pm

This is not uncommon to create an ODS as a buffer between your source and your data warehouse. I am sure Ken can give the long answer why this is important.

DSguru2B · Post by **DSguru2B** » Tue Mar 27, 2007 8:56 pm

I agree with Kim. Not uncommon at all. I implemented something along those lines at one of my client's. Ken can definately throw some more light on it. We all love his in depth answers.

ray.wurlod · Post by **ray.wurlod** » Tue Mar 27, 2007 9:51 pm

A great start point is developing a business glossary - centralizing (and getting agreement upon) the definitions of business terminology. Only when you have this in place, and agreed to, does it make any sense to go forward with designing a centralized model.

I suggest that you look at IBM's business glossary offering, as it can be stored in the same location as metadata used by modelling tools, ETL design tools and metadata analytical tools, and can be leveraged by all of them.

RodBarnes · Post by **RodBarnes** » Wed Mar 28, 2007 8:56 am

Thanks for the replies. Sounds like I am in the right direction. I look forward to Ken's input.

Ray: We are actually looking at the IBM Business Glossary and Metadata Workbench. Very impresive capability. The challenge (of course) is getting management and IBM to agree on a price.

DSguru2B · Post by **DSguru2B** » Wed Mar 28, 2007 9:05 am

Is'nt that always the challenge

Good Luck.

ray.wurlod · Post by **ray.wurlod** » Wed Mar 28, 2007 12:31 pm

Wait for the challenge in getting all the silo owners to agree on consistent enterprise-wide terminology. How good are your negotiation and dispute resolution skills?!