The answer, unfortunately, is that there is no good answer.
I use old PCs as storage servers. The flexibility is nice, but they use a lot of power and output a lot of heat. Also, when you stick a PC in a closet and forget about it, you’re not going to notice when the CPU fan siezes (and eventually it will sieze). Your first inkling that something is wrong is going to be when the CPU has overheated, killed itself, and welded itself into the CPU socket. It’s possible to go with a fanless setup, but then you’re likely talking about something purpose-built, and not something you just have lying around.
Stuff like this seems neat at first, and certainly addresses power and heat concerns, but has other problems. First, firmware quality varies greatly, even among different generations of a device with the same model number. It’s easy to get stuck with something that needs power cycling every few days to clear up firmware crashes. Second, almost all of the low end devices use USB 2.0 as the interface to the hard drive. If you have a switched, Gigabit Ethernet network and even a couple of heavy users, USB 2.0 quickly becomes a performance bottleneck. Finally, most of these devices force you to use FAT32 as the filesystem on the attached drive. FAT32 limits maximum file size to 2GB, limits maximum volume size to 1TB, and is a recipe for data loss in the event of a power outage.
I don’t have a lot of experience with higher end devices because I’m poor and they are well outside my price range. They tend to use proprietary filesystems that eliminate the drawbacks FAT32 gives you, but prevent reading your disks in any other device should your device die. As others have said, RAID is really more of an uptime solution than a data security solution. The other thing to keep in mind is that if the probability of failure with a single drive is x, then the probability of failure with three or five drives is 3x or 5x. Sure, RAID will keep your data intact, but you’ve still got the administrative hassle of ordering a replacement drive, installing it, and running a rebuild.
My solution to date has been an old PC doing NAS duty, and a bunch of USB drive enclosures. I put identical drives in the PC and all of the drive enclosures, and run a full backup from the PC to one of the USB drives every few days. Failure of one drive isn’t a huge deal and I can take my time ordering and setting up a new one. It is far from perfect. There’s a lot more administrative hassle than I’d like, and as electrical rates continue to rise the cost of keeping the PC running is becoming more noticable.
I’ve often thought of starting a business that does nothing more than install, monitor, maintain, and back up home servers and gateways. I bet there are a lot of people out there that would pay $20 or $30 a month not to have to think about this crap.