« The Smoking Gun | Main | I've looked at Cloud from Both Sides Now »

May 04, 2009


Feed You can follow this conversation by subscribing to the comment feed for this post.

Shou Kuo

I assume by "old-school" you are referring to the "traditional" data warehousing architecture which consists of ETL tools and "common" data model for the warehouse. Is the inability to deliver having to do with the semantic deficiency of ETL and/or common data model? Does your approach take advantage of the SOA: Canonical Data Model?

See: http://kjellsj.blogspot.com/2007/07/soa-canonical-data-model.html

Lyn Robison

A canonical data model is certainly important for understanding the big-picture realities. I think the trick is to figure out how to build a CDM without boiling the ocean. A good approach is to start with one important business topic for which there is clear business justification, model it, and then expand from there.

Shou Kuo

The referenced post above seems to also advocate an approach of starting the modeling process by focusing on a well-defined business domain and related processes, rather than databases and other technology. In this case, the CDM seems to serve the purpose of a "conceptual" data model for data integration across multiple disparate systems, which could be implemented using a standard language such as the W3C XML Schema. As a result, there would be no need to develop the corresponding "logical" and "physical" data models if the "target application" for data integration is also implemented using XML-based technologies rather than the "traditional" relational or object-oriented technologies. Does this make sense?

The comments to this entry are closed.

  • Burton Group Free Resources Stay Connected Stay Connected Stay Connected Stay Connected

Blog powered by Typepad