As you may know, I love backup hardware, software, and media; and I have also been looking at a backup system for my "very large" home virtualization computer. Since I started looking, I went from "oh, I'll use blu-ray disks" to DAT tapes, back down to RDX cartridges, and eventually found that LTO was most likely to be the most cost-effective solution, especially as I start to fill the ten-terabyte data store I've got. (Most of that space is media and other archives, which I am sharing with my housemates, along with my photography collection, and I am likely to be storing backups of my laptops on the machine as well.)
I've been taking a look at the backup situation again, and I've actually discovered some new stuff (which I'll post about later) that means I may be making progress on the backup situation for TECT.
LTO6 is fairly exciting for me for a few reasons: The first is that it increases the capacity per tape pretty significantly, to 2.5 terabytes (native), and increases the speed at which data moves to the tape to 200MB/second, which is a good clip faster than the fifth generation LTO technology.
The biggest things this means for TECT, presuming LTO6 drives are anything resembling "affortable" is that I can back up TECT onto fewer tapes in less time. With a backup array that uses reasonable disks and a reasonable controller, meeting the needs of an LTO 6 mechanism shouldn't even be too wildly difficult, although even with LTO 6 out, I may consider using LTO5 if I can't get a reasonably cost effective array that can deliver the speed that 6 needs.
Even though it's a little bit ridiculous, I am kind of still waiting for the 8th generation DAT/DDS product to come out. "DAT 640" as I'm going to call it now should skew the DAT line back down a bit in my graph of data storage technology cost, and be fast enough that over the course of a week, I should be able to move a whole TECT backup to tapes. The biggest thing to worry about is whether or not I can physically carry the number of DAT tapes it would take to back up a ten terabyte dataset. (That's 63 tapes, presuming no compression can be achieved, and DAT320 tapes, which have a native capacity of 160GB.)
I also still look in on RDX from time to time. RDX intrigues me a lot because it's more reasonable to use without a large staging area, and in the beginning, for a single server or workstation, it's a lot more incremental, cost-wise. Unfortunate things about RDX include its unsuitability as an archive media, the lack of an auto-loader that behaves in any kind of predictable way, and even if 10TB worth of media can be bought, what about that second 10TB in order to provide an off-site copy? What about the copy I want to keep for six months? It's just not very viable for what I'm doing with TECT. (Although that's not stopping me from running the numbers every so often.)