HomeSQLServerPedia SyndicationSSAS: One giant cube or many small ones?

Comments

SSAS: One giant cube or many small ones? — 9 Comments

  1. excellent post James. keep it up.
    by the way, regarding the point you mentioned in disadvantage of multiple cubes i.e. “Requires a SSAS project in BIDS for each cube”. – BIDS allows us to have multiple cubes in 1 single SSAS project. and if we choose this method, then its beneficial for developers. But if we have cubes in different SSAS projects, then its too much to develop and maintain just like you have explained.
    thanks, khilit

  2. Pingback:Microsoft SSAS Solution: Cube Design Considerations | Khilit's Business Intelligence Blog

  3. Nice article James. I must admit the choise bewteen the two approaches is quite often a difficult one as no single option is clearly better.

    Another way of looking at this issue is from a business usage perspective. From my experience the “one giant cube” approach is better for enterprise level reporting i.e. if you want to see what’s going on across the organisation and across business processes then this is better. On the other hand if the reporting is across discrete business units, departments or business processes then the many small cubes approach may be better.

    In my opinion it is better to consider the business scenario before the technical implications i.e. when trying to decide whether to have one or many cubes.

    Keep up the good work!

  4. Great article.

    As we are after SQL Server 2012 release, It might be worth clarifying it is for previous versions or SSAS 2012 Multidimensional. Some of the advantages don’t apply to SSAS 2012 Tabular (processing dependencies) but I suspect we cannot call tabular a cube (although I do when I speak with business people as they don’t care and generally it is easier for them understand 😉

    I also assume when you refer to database you always mean SSAS Database not data warehouse database? Is that the case?

    One question I have are perspectives and linked cubes. Any thoughts on those two features?

    Take care and thanks for sharing your thoughts with everyone 😉
    Emil

    • Hi Emil,

      Glad you like the article. You are correct in that this is dealing with SSAS multidimensional cubes. And yes I am referring to an SSAS database.

      Perspectives work great at limiting what a user sees, especially if you have a big cube. I don’t link linked cubes since linking fact tables is not much help since linked fact tables can only be used on linked dimensions, plus changing the structure breaks the link.

      • Hi James,

        I have similar experience. I like perspective also because first can be physical names and second can be friendly name so rename of friendly name won’t break existing MDX.

        I’ve seen linked cubes being used but I didn’t like it for the reasons you specified, I think it is much simpler to maintain without dependencies.

        Take care
        Emil

  5. Pingback:Bus Matrix Presentation Follow Up Q&A « byoBI.com

  6. Dear James,
    We have around 30+ small cubes in our BI project and would like to know if there is any way to deploy all these 30 cubes at once through a script or any other process.
    Currently we are deploying each bim file individually which is time consuming task.
    Many Thanks,
    CR

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>