Larry's first paragraph summarized it up perfectly.
It case it makes a difference, we have 52 disks, 4 of these are SSD's, the
rest are all 140's. They're all in 5886's. The SSD's have their own. The
disk unit preference on the BPCS data is set to the SSD's.
And some may question measuring performance on month end processes.
However if the problem you're trying to solve is "How can we reduce data
outage by reducing the length of month end?" then, by God, that's the
problem you focus on. Apparently the people with the checkbook thought