Basically the point we need to get across is that you receive a drive and you can put data on it however you want. We provide a helper for you, but if you want you can use whatever backup mechanism you'd like. We just take the data and sync it to Amazon Glacier then resecure the drive.
I hope you have a lot going/planned in the UX part and would love to hear about it :-).
An image of the drive? Files on the drive? More data would be helpful.
Do we ask for a particular drive size and you send it?
What if I've got a 2TB drive with TrueCrypt volumes and I just dd it over to the drive you provide? Do you put the entire drive image on Amazon Glacier?
Or are you expecting a readable filesystem?
More importantly you also need to put some contact info on the site and not use "whois privacy" for your domain name. How is anybody supposed to know who they are dealing with and why they should trust you?
I mean seriously why in the world would anyone send someone who is (virtually anonymous) a copy of their hard drive (not that putting contact info alleviates my fears or can't be faked) but it is certainly a step in the right direction.
As a social engineering test I wonder what the results would be if some official looking company sent 100 hard drives out to random small companies along with software to clone the hard drive and a postage paid envelope (forgetting mail fraud etc issues or with an overseas address)
Thanks again to everyone!
http://aws.amazon.com/glacier/faqs/#How_will_I_be_charged_wh...
Also, be aware that you'll not be able to de-dupe files across users (this also has issues if you could) as you're encrypted. Also, if the encryption key is different, you won't even be able to do this cross backups...which will suck.
1. What kind of drive will you send around? Spinning disk? How much capacity?
2. Is there a special software that customers have to use or could I dd if=/dev/sda of=/dev/sdb?
3. In case a full clone is okay (see question 2), which filesystems are supported when I want to retrieve a single file online?
Also, here are two observations:
1. It occurs to me that the time for a full backup is much higher than an incremental backup. You should probably cover on the page whether you support incremental backups or not. Having a slow computer for many hours could be a deal-breaker.
2. The claim about other backup systems trying to force some kind of organization scheme on you (can’t quote since the page seems down) is vague. I have used various backup solutions (bacula, obnam, custom rsync scripts, bup) and none of them has forced me to do any re-organization of my data.