Memory management of powerpivot workbooks in sharepoint

Question

I wonder how the memory management in Sharepoint handles PowerPivot workbooks.

For example:
A workbook with the PowerPivot-Add in in Excel with about 1 GB data, published to Sharepoint

After I published the workbook, ten users open the workbook via excel services. I think that there are two possibilities to handle this.

1st possibility:
Analysis services stores the data in the VertiPaq engine and excel services is only a kind of frontend. => The total memory needed is only about 1GB no matter if ten or one users are loading the workbook

2nd possibility:
Analysis services loads the data with the VertiPaq engine to deliver the 1GB to each workbook which is loaded. So for each workbook a copy of whole 1GB data is created. => The total memory needed depends strongly on how many users load the workbook. (1 user : 1 GB; 2 users : 2 GB; ... 10 users : 10 GB)

Thanks a lot for any hints

J.

Answer 1

The PowerPivot workbook is loaded inside Analysis Services running in SharePoint integration as a AS database, and this AS instance will service the client requests targeted for this PowerPivot workbook for the entire farm. So from this point on, it is just like the client server architecture in the traditional Analysis Services system in terms of memory management.

Thanks,

Lisa

Answer 2

The PowerPivot workbook is loaded inside Analysis Services running in SharePoint integration as a AS database, and this AS instance will service the client requests targeted for this PowerPivot workbook for the entire farm. So from this point on, it is just like the client server architecture in the traditional Analysis Services system in terms of memory management.

Thanks,

Lisa

Answer 3

Hi Lisa,

thank you very much for your reply. I am not sure if I understand you in the right way. Does this mean that only the query results are deliverd to the client and not the whole database (in the powerpivot workbook) ?

best regards

Joachim

Answer 4

Hi, Joachim,

In the case of web browser requesting the PowerPivot Excel file published to the SharePoint farm, only the query result is returned to the client (Excel Calculation Services) and Excel Services renders the result in Pivottable or PivotChart with the same fidelity as in the regular Excel desktop client.

If you choose to open the published file in regular Excel running on the client machine, the embeded PowerPivot database will be loaded into the regular Excel process when user clicks the PowerPivot ribbon to launch the PowerPivot environment inside Excel. The memory consumed by the regular Excel process is on client machine only, not part of the SharePoint farm.

Thanks,

Lisa

Answer 5

Thanks!

Joachim