There are two major benefits to hardware RAID:
- Performance. If that's not a consideration, don't consider it.
- Battery-Backed Cache.
2 really goes hand-in-hand with 1, but it also plays a role from a data integrity standpoint: Let's consider two identical machines, one with Software RAID, one with a battery-backed hardware RAID, both with write cacheing enabled.
If we write a whole bunch of data to these two machines, then rip the plug out of the wall before the data is sync'd to disk (by the OS on the software RAID box, or the controller on the hardware RAID box), we now have two different scenarios:
The Hardware RAID machine will, when power is restored, see that there are writes in the cache, spin up the drives, sync the data to disk and move on with its life.
The Software RAID machine will, when power is restored, reboot normally. Since the system RAM lost power the writes that were stacked up waiting to be flushed out to disk are gone forever. The results here could range from inconvenient (a little data loss) to catastrophic (the partition table has been eaten, vital irreplaceable data has been left in a half-written, corrupted state, etc.).
Now that I have portrayed the nightmare scenario, consider the likelihood of it happening. Presumably your server is on a UPS, ideally with redundant power supplies fed from different UPS circuits. The chance of a catastrophic failure in that scenario is relatively low, and you may be fine trusting software RAID here.
In exchange for taking that (theoretically small) risk you aren't beholden to a particular brand of controller and revision of the controller firmware in the event something goes horribly wrong, and you have the flexibility afforded you by software RAID (which often includes mirroring only part of a drive, or being able to mirror across drives of different sizes, spindle speeds, etc. -- Possibly a consideration if you have Frankenstein hardware).