locked
Managing custom and core BI Cubes RRS feed

  • Question

  • Hi,

    I want to seperate out core and custom BI cubes so that for a product I can have some generic schema, SSIS packages and Cube and for clients I can give custom schema, packages on top of core. Can anyone tell me how to achieve this? So may be one person will keep on developing the core BI solution and any updates / corrections to core should be merged into the custom solution. is this possible?

    THanks,

    Ameya

     


    • Edited by Ameyask Wednesday, December 28, 2011 10:16 AM
    Wednesday, December 28, 2011 10:15 AM

Answers

  • Hi Ameya,

    I think it is a best practice to have single data model for entire warehouse. For requirements which you mentioned above, should be managed with data source views in the cube. If you dont have specific column which you are looking for in the model, you have to add it to the model first and derive it at the cube level. If we have multiple copies of data models each catering for specific requirement, then maintenance becomes tedious.

    Is there any specific reason why those new measures are not being added to the warehouse instead of creating separate schema? Please let me know.

    Please correct me if I got your requirement wrong.

    Regards,

     


    Phani Note: Please vote/mark the post as answered if it answers your question/helps to solve your problem.
    • Marked as answer by Challen Fu Thursday, January 5, 2012 10:01 AM
    Monday, January 2, 2012 1:00 PM

All replies

  • Hi Ameyask,

    What's the meaning of "for clients I can give custom schema, packages on top of core"? If you want to customize the cube based on the different customer, I think you can create a different cube based on the same database dimension collections, or you can make use of Perspective. Actually, Cubes can be very complex objects for users to explore. A single cube can represent the contents of a complete data warehouse, with multiple measure groups in a cube representing multiple fact tables, and multiple dimensions based on multiple dimension tables. Such a cube can be very complex and powerful, but daunting to users who may only need to interact with a small part of the cube in order to satisfy their business intelligence and reporting requirements. At this time, you can use a perspective to reduce the perceived complexity of a cube in Analysis Services. A perspective defines a viewable subset of a cube that provides focused, business-specific or application-specific viewpoints on the cube.

    For more information about Perspectives, please see: http://technet.microsoft.com/en-us/library/ms175338.aspx 

    If I misunderstand you, please feel free to let us know.

    Thanks,
    Challen Fu


    Please remember to mark the replies as answers if they help and unmark them if they provide no help.

    Monday, January 2, 2012 8:16 AM
  • Hi Challen Fu, Thanks for your feedback. Let's take a following example. i Have a core BI / Core cube solution of my product which has a fixed schema and we give to our clients. let's take an example of a typical sales db. I have Items list as a master list / dimensions and the measures hold the daily sales data. Now, for a client we need to make some additions / modifications to the same so we created a schema where we added some more measure tables and wrote reports on this. Now my core BI team kept developing the core cube and its shcema; and now I have a few things in the standard project which I wish to copy to the custome BI/Cube project. Can we achieve this or we have keep making the similar changes at both locations in order to keep them in sync? Let me know if you need more information. Thanks, Ameya
    Monday, January 2, 2012 9:51 AM
  • Hi Ameya,

    I think it is a best practice to have single data model for entire warehouse. For requirements which you mentioned above, should be managed with data source views in the cube. If you dont have specific column which you are looking for in the model, you have to add it to the model first and derive it at the cube level. If we have multiple copies of data models each catering for specific requirement, then maintenance becomes tedious.

    Is there any specific reason why those new measures are not being added to the warehouse instead of creating separate schema? Please let me know.

    Please correct me if I got your requirement wrong.

    Regards,

     


    Phani Note: Please vote/mark the post as answered if it answers your question/helps to solve your problem.
    • Marked as answer by Challen Fu Thursday, January 5, 2012 10:01 AM
    Monday, January 2, 2012 1:00 PM