The source system is available for upload in approximately 1-4 hour windows throughout the day. The power supply is cached and requires a refresh at these intervals for about an hour I would guess. I haven't had time to benchmark this since the source system arrived only yesterday.
Do these managers have pointy hair, by any chance? How big are the rows? Can TPump, running standalone, achieve these benchmarks? (If not, then clearly DataStage driving TPump will not be able to. Don't sign up for anything doomed to failure - push back against unrealisitic demands - the more acceptable term for this is "expectation management".) Make your own benchmark, using two-byte rows, just to illustrate how pointless their requirement is.
Can't you convince them that MultiLoad sounds cool too?
You certainly seem to have enough AMPs, provided they're not being used for other things.
IBM Software Services Group
Any contribution to this forum is my own opinion and does not necessarily reflect any position that IBM may hold.
Having just delivered an app using TPump as the load utility, I can assure you that TPump WILL NOT be the fastest solution. TPump's purpose is essentially to allow trickle feeds. We used TPump for our lower volume tables (to satisfy the pointy haired boss's requirement for 'cool') and used MLoad for the heavy lifting.