HomeData warehouseOperational Data Store (ODS) Defined

Comments

Operational Data Store (ODS) Defined — 10 Comments

  1. I liked your second sentence.

    15 years ago I worked on a Big Five system where the reports ran, slowly, on the OLTP db. I put together a nice project charter proposing that we build a data warehouse. Denied!!! “DW project failure rates are too high, you must come up with another solution”. I opened MSWord and did a “find” for “warehouse” and a “replace” with “mart” and resubmitted. Denied!!! “Although dm project failure rates are less than dw failure rates, it is still too risky. You must come up with another solution.” Back to MS Word…this time I did find on “data mart” and a replace with “ODS”. I resubmitted. Approved!!!

    Eight months later I was giving a project closure presentation where I outlined the final architecture of the “ODS” as well as how the reports now ran in minutes instead of DAYS.

    One of the project sponsors stopped me and asked, “You just built a data warehouse, didn’t you.”

    Moral of the story: a rose by any other name…

    • Hi Dave,

      Do you have a model copy of your proposal. I have similar to your situation and team assigned me for doing the paper work

  2. I agree with your definition of an ODS and it’s value. On a number of projects I’ve worked on, there has been a virtual prohibition of even mentioning the term ODS. I’ve used the global search and replace method to change it to Reporting Data Store and they loved it.

    Any idea where the ODS got the reputation it has in so many companies?

    • Thanks for the comment Mel! I don’t know where ODS got the bad name. Honestly I have not run into a client yet who gave me dirty looks when I talked about an ODS 🙂

  3. I think ODS has a bad rap because it’s usually implemented at companies that are poorly managed. If enough time is spent in the planning stage you are more likely to successfully build a data warehouse. Too many companies try to build a data warehouse, but end up with an ODS because they didn’t plan or put much effort in gathering requirements. In many cases the ODS is an appropriate need, but often it’s the result of poor project management.

  4. Pingback:Data Loading performance considerations with Clustered Columnstore indexes | Garrett Edmondson

  5. We do have Operational Data Store implemented on SQL Server with near real time reporting, and we build the ODS because of the similar reasons that that you mentioned. We have too many sources where entire business process runs and its imperative for us provide ODS to give the operational reporting.

    Coming to the question: What’s your opinion on building ODS using big data technology. Due to the kind of integration and transformation we have, i feel, big data may not be performant and cost effective. Any thoughts?

  6. Great explanation on the differences between the ODS and data warehouse. Totally new to warehousing and this has shed some light for me. Question on the ODS layer, in a real world scenario, which is better practice, full load or delta?

Leave a Reply

Your email address will not be published. Required fields are marked *

HTML tags allowed in your comment: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>