Its important to note that Big Data and NoSQL are geared toward unstructured data and both Kimball and Inmon are driven off of structuring data into some form – either 3rd normal or star/snowflake schema.
The other thing I would add to the discussion is the segregation of the two functions of BI – analysis and reporting.
Both Inmon and Kimball, I believe, ultimately fail in the true analysis function of BI because by the nature of strucuturing data, you are already dictating a relationship between data that may or may not lead to factual causation and actionable information.
Kimball and Inmon due extremely well in providing metrics based upon presubscribed relationships between the data which works well for first level analysis and reporting.
For deep dive analytics, technology has come far enough with NoDQL, Hadoop, MapReduce, etc, that we no longer need to structure data in an Inmon or Kimball model.
So the answer of are the Kimball/Inmon models for warehousing obsolete depends on the function within BI.
Reporting – no, analytics – yes.
More Resources on What are the relative strengths and weaknesses of the Bill Inmon and Ralph Kimball approaches to data warehouses in 2012:
http://www.dbms2.com/2010/10/06/ebay-followup-greenplum-out-teradata-10-petabytes-hadoop-has-some-value-and-more/
Read more about What are the relative strengths and weaknesses of the Bill Inmon and Ralph Kimball approaches to data warehouses in 2012:
146 Policies and Procedures Ready to Use – Best Practices in IT Policies and Procedures
This question originally appeared on Quora.