
Netgear/Infrant has never gone into the specifics of how it's done, but I'm guessing the drives are partitioned and the partitions are then RAIDed to ensure drive-level failure can't cause a problem. I know I've seen people do the same thing in software on x86 machines (in LVM, maybe?), so I'd guess that's what they're up to.
That's exactly how it's done. On my 1000S with v3 firmware, the raid partitions were sd[abcd]3 and were in raid 5 with one big lvm volume on top. Unfortunately, I know this because I had to recover my data this way after my 1000S went south due to firmware image corruption. Getting the raid reassembled under linux was easy: modprobe md, force reassembly of the raid, scan and activate the lvm volume group, mount the volume.
Netgear support was next to useless when reflashing the CF didn't solve the problem. I wouldn't call their tech support terrible; I figure that they probably solve at least 95% of their customers' issues. I suspect that they could have solved my issue as well, but once you're out of the warranty period, they really have no motivation to do so. The 1000S was only warrantied for a pitiful 1 year; at least the current versions come with a very respectable 5 year warranty.
The other sticky thing about purchasing these devices from any company is that while you may care about your data, they don't. Their responsibility within the warranty period is only to keep your hardware operational. Backups are wonderful, and presumably everyone will back up the super important stuff at regular intervals, but most people don't have another place on their network that will hold the full 1.5->4+ TB of data that these things can store. Sure, you could buy additional units/drives/whatever, but that's pricy if what you're protecting is just music/video files or something of that nature.
Honesty is for the most part less profitable than dishonesty. -- Plato