Are you planning on running through a minefield with only your hands over your ears as protection, or do you want this with maximum resilience?
Maximum resilience needs two identical computer rigs & a method of switching on the fly when one goes down.
Note I didn't say if.
Assuming that's beyond budget, I would change direction & look at a USB interface that can do it in the box.
Line 6 make mic preamps like the Pod Studio UX 2 & UX 8 that can in real-time feed pre-programmed effects straight back out of the output with no computer latency.
Stenberg/Yamaha have just released something similar, the UR-RT series The dspMixFX chip can even be controlled from your iPhone [but I've never used it, so idk how workable that would actually be in practise]
They are both controlled by the computer, but the effects are in the external unit itself.
Neither is going to be able to do multi-band compression on the fly - but if you were to do something complex like that through the computer, you'd be looking at 100ms latency - completely unacceptable.
I think the hardware i/o might be the middle ground.
Alternatively, a dedicated hardware unit that can do it without computer control - but I couldn't name one off the top of my head for vocals.