Global DataWarehousing Model

[Former Member]

My client wants a Global Data Warehouse(Americas,Europe,Asia etc). with the capability of reporting at the global level and drilling down to the regional and sub-regional levels.

What would be the advantages and disadvantages of the following two approaches:

1.Approach 1:Build the warehouse with one InfoArea, One InfoCube etc. per area with all the users from Europe,Asia and America accessing a single InfoCube.

2.Approach 2:Build separate Infoarea for Europe,Asia and America with users accessing the reports at their regional level and create a multiprovider for users who want to access the global reports.

Any insight and experience would be helpful.

Thanks in advance.

[Former Member · Accepted solution]

Second approach is ok if you are looking for implementing warehouse for small number of locations, I mean it will be difficult if you have 30 locations and building a 30 cubes for a same business solution.

Identify your business areas (like billing, pricing, compensation or benefits etc. and then prepare a matrix and see how many combinations can possible? If the numbers are within the limit then you can go for this approach otherwise you can handle this with less number of objects and do the regional reporting based on Authorization

Hope this helps

[Former Member · Accepted solution]

Hi Ashish & Dinesh,

Thanks for the reply.We are looking to build infoareas only at the regional level(read as Americas,Europe and Asia-and not 30 locations separately)and then feed it into Global Multiprovider to meet the global reporting requirements.

The reason for that would be because of feeding the data at the end of the day for each region.(Diff. process chains)

What are your opinions about the maintenance and flexibility of the two approaches?.

Thanks in advance.

[Former Member]

Hi,

You could probably have a regional ODS and Cube. Now these regional Cube can then feed the Global Cube with data from all the regions. This way the management can report on the global data and then drill down to the regional ODS/cube.

Bye

Dinesh