It’s time to start using Amazon Glacier Deep Archive for backups

Let’s talk data storage and cloud backups. It’s time: I’ve started uploading backups to Amazon Glacier Deep Archive, which is about $1/TB-month.

My data is categorized as “warm” and “frozen”. “Warm” data is active: it is kept on my workstation or on my primary NAS, and is backed-up to a second local NAS as frequent, versioned snapshots. Both NAS boxes are multi-drive redundant, and my local workstation data is continuously backed up to the primary NAS using ResilioSync (and pushed to the second NAS in snapshots). I’m setting up a third, offsite NAS to receive snapshots, as well (friends with fiber!).

“Frozen” content is locked, and has been organized into labeled, 12-TB chunks, which are stored in triplicate:

• 1 copy online (read-only) on my primary NAS (QNAP, 8 x 12TB drives, 10G Ethernet). I can access this copy quickly, but the volumes are enforced as being read-only.
• 2 additional copies as offline disks (1, offsite).

“Frozen” data isn’t actively backed up.

I now plan to store one of the copies of the “frozen” data on Amazon Glacier Deep Archive, and keep only 2 copies locally. The chances that both of my local copies will be simultaneously lost is slim, so it is unlikely that I’ll need to initiate a data retrieval from Amazon (which only takes 12 hours to kick off, it’s so not a huge deal).

At the moment, I have about 36TB of “frozen” data, so it will cost me ~$36/month to store it on Amazon Glacier Deep Archive. I’m amazed that this sort of reliable cloud storage can be so cheap! Also, since it can be accessed as standard S3 (via needing to relay via S3, which is how the non-deep-archive version of Glacier), it’s trivial to dump data onto it. I’m uploading at 115MB/s (big B!) via AT&T residential fiber. This will be a great move for insurance and peace of mind.