Daniel Graham of Teradata spoke Tuesday evening at the SDForum BI SIG on the topic “Data Management in the Cloud.” See my article summarizing his presentation. It’s interesting to compare his presentation with that of Donovan Schneider from Salesforce.com, who spoke at the BI SIG a couple of months ago. (See article on his presentation here.) Graham described how to move one or all of the BI steps of extract, load, transform, and analyze to the cloud. Schneider had described a more radical approach whereby data is kept in the OLTP system and analyzed real time. I could see how each of these approaches would make sense in different scenarios. Both presenters agreed on one point: the cloud is ready to handle BI for scale to medium-sized implementations, but not for massive data sets.