Hey Kevin,

Working on PowerPivot via oData, I have some interest in your findings. So, if you got a minute...

What does database size have to do with the decision to upsize or not upsize the data?
What advantages are gained in PowerPivot by upsizing to SQL Server? Or, what factors should be considered in relation to PowerPivot and the backend storage?


Hi, Craig,

I've haven't timed it down to the second, but I've found that the import (and refreshes) of large data from SQL Server to PowerPivot is much faster than reading from other feeds/sources.

It's certainly not a requirement to go against SQL server tables - many demos show the entire mash-up concept - and many "one and done" jobs wouldn't benefit from upsizing.

But for repeated instances and where the resulting output will be integrated with SharePoint, IMO using SQL as a source/staging area is probably the direction that many companies will go - especially in places where PowerPivot will supplement existing DW systems.

Personal opinion, I find the process of going directly against SQL Server to be a little easier for end users (the target consumers for PowerPivot) than from other sources. But obviously that's not the entire audience.

I really like PowerPivot, but my concern about PowerPivot, in any sharing / SharePoint integration, is the size of what's being deployed. Even with Vertipaq, there's potential to slow down an environment. Even when I take a portion of AdventureWorks, create a few pivot charts, push it up to my local sharepoint site on the same box, and then use it in a powerpivot gallery, there's lag at various stages, and that's just running single user. I think MS created a great tool in PowerPivot, and the new viewer that got announced at Tech-Ed is pretty slick - the next year will be interesting to hear where the implementation challenges are.


re: OData, I'm assuming you've seen some of the things Beth Massi has written over the last few months, she has some good content.