Member's data mining project

I?m in the middle of putting together a corporate data warehouse for a telecom company. This year, the company has planned deployment of data mining functions.
What are some key points related to architecture?
In our environment, we are based in IBM DB2 and IBM DB2 OLAP, running on top of RS/6000 (MPP architecture) hardware.
For this reason, we'd like to know which products can be integrated in this parallel environment, with extractions running in parallel, processing running in parallel, and if it's possible, loading data from the data warehouse database in parallel, without constraining the processes in extracting files to load into the data mining tool's repository.