Data Vault, Kimball Star Schema, Inmon CIF

i’ve been asked, over and over and over again throughout the years to define the differences or compare and contrast the data vault with kimball star schema or kimball warehouse, and inmon cif.  well, let’s see if we can set the record straight here.

1) the data vault is not a framework, it is a two part implementation standard.  the first part is a set of data modeling rules for implementing the data model portion of your data warehousing project.  the second part is *optional* but is a project implementation/project plan for implementing your data warehousing project.   (you can read more about each of these parts in subsequent posts).

2) the cif (corporate information factory) is really a framework, much like the zachman framework – only the cif framework focuses on data warehousing components and overall architectural slots. 

3) the kimball warehouse “architecture” is a framework, some have called it kimball bus architecture, it also (like the cif) focuses on data warehousing components and systems design.

neither of these two frameworks are “competitive” in nature to the data vault, however the cif framework naturally fits better, because the data vault requests that you build a three tier setup for scalability: staging, data warehouse, and data marts/release area.  however, if you really are keen on kimball bus architecture, you *could* concievably build the data vault model as your kimball staging area – although no-one i know of has followed this route.

4) the kimball star schema – is a data modeling technique which is different than the data vault modeling techniques.  don’t confuse the data vault modeling techniques with the methdology components please.  you can and should compare and contrast the kimball star schema with the data vault modeling techniques, this is a valid claim – and yes, there are differences, and yes there are pros and cons.  some of which i will address in future blog entries.

so once again: 

  • cif & kimball bus = architectural frameworks (don’t tell you how to implement). 
  • data vault methodology = project plan + people + it workflow (tells you how to implement). 
  • data vault model & star schema = data modeling techniques (tell you how and what the rules are to modeling your enterprise data warehouse).

i hope this helps clear up most of the confusion

Tags: , , , ,

No comments yet.

Leave a Reply