I understand what you mean about complexity. My first _large_ raid array was at sgi. We had 2 cabinets worth of drives on 4 arrays. Probably like 60 drives in _one_ array with 12 controllers per array (5 drives on each controller). 2 of the arrays were mirrors of the other 2. Some complex **** there. It was a pain trying to figure out which drive failed (until I figured out how to use the "blink the lights" utility...) And fail they did. At least one a month. That is where I learned about striping and stripe size calculations. What works for the OS build group does not work for the streaming media labs...