Quote Originally Posted by msa1500
Mean Time Between Failure :

It all has something to do with statistics , this number gives you an idea how long a component will function without failure.
But theres more :

Lets say i have a disk with a mtbf of 1000000 hours.
so when i hook up one such disk theres a high probablity that the disk will fail within 1 mil hours. (99.8%). But when i hook up 200 disks within one array. The array will have a mtbf of 1000000/200 = 5000 hours , so there's a high probability that one disk will fail within 5000 hours of operation (99.8 % )

This is all to give you an impression what this mtbf means.
It doesnt mean that a disk will actually fail within 5000 hours.

Greetz
I know you're a professional storage guy, but IDE (P-ATA and S-ATA) are so common to run 24/7 these days. 99% of the home users who have a box running 24/7 use IDE and lots of the smaller and mid-size companies heavilly depend on IDE as well. I am aware of the MBTF term and what it means, but I've seen more then my share of (nearly brand-new) SCSI harddisks failures. Especially accounting the 'fact' that SCSI is professional and IDE disks are not...

Anyway, a S-ATA RAID-5 array using 4 disks provides (on a small network) the same performance with much larger capacities for a significant lower price with practically the same data integrity ratio.

So SCSI is hardly an interesting product here. Unless you can deliver me a 1 TB SCSI NAS device for EUR. 800,-- ofcourse