Joining data from 2 cubes - Multiprovider or Infoset?

Hello,

We are just underway with our BW 3.1 implementation, and are starting to evaulate methods used to combine data from two related cubes.

In reading the help for Multiproviders, it says that the Multiprovider is utilizing a union type join, and Infoset's utilize a table intersection join.

Is this the proper thought process for addressing a need to combine two cubes that have a relationship (ie A document no. = B document no.)? I'm sure this is a simple process, but any guidance on an approach to this would be helpful.

Thanks

Bill Coombs