That was pretty much my understanding of it after taking that course, and I think that must have been what they wanted us to take away from it, seeing as I got a ridiculously high score for my exam paper for saying as much.
That case study is interesting though, your post on mainframes reminded me about it the other day: old mainframe system running 420k transactions per night, shiny new erp system running on shiny new hp servers managed 10k transactions per night.