SSIS: Experimenting with the target cache lookup

As a part of the project I’m currently working on, the development team is very enamored with the cache transform in SSIS. I consider myself pretty experienced with SSIS, and I’ve always stayed away components that come anywhere close to touching disk with SSIS, as I believe it slows down package performance. The cache transform component takes a query that you want (what you’re looking up) and saves it into a .caw file for SSIS to use in other locations. This is an alternative to the in memory record set destination as the result of transforms as well. As a use case, let’s say you have a customer dimension that every single fact table you load wants to reference. In the Kimball dimensional world, this would be your conformed dimensions. Do you really want to do the lookup on every package to that customer table, or would you rather save…