21 May
2007
21 May
'07
6:46 a.m.
For data backup, it's a false economy to compare
the cost of the
backup media to the cost of other backup media or to the cost of gas
or to the cost of tea in China.
The true comparison is the cost of making a backup vs
the cost of not
making a backup.
Not really - if you've decided to do backups (or actually even if
you're only considering doing backups), there is nothing wrong with
trying to find the cheapest way consistent with your particular
environment's threat model and risk tolerance levels.
And *that* is where comparing the cost of backup media to the cost of
other backup media comes in. (For example, it's entirely plausible a
site would decide "if we can get media for under $0.01/GB then it's
worth backing this up, otherwise forget it, we'll just recreate it if
we need to".)
For more than a decade folks have argued about the
cost of backup
media vs another backup media, and ignored the cost vs value of doing
the backup at all.
Yes, this is..suboptimal. But just because you shouldn't ignore thing
#1 doesn't mean that ignoring thing #2 instead is right.
The cost of the media has been in the noise compared
to the effort
involved for a long long time, at least for those of us in the West.
I disagree.
At work, for example, we write (some of our) backups to real tapes.
I'm the person who usually changes the tapes. I just did a
back-of-the-envelope estimate, and I figure it costs us about a dollar
each time I have to change a tape.
So, positing the existence of tapes small enough that a run took two
(it currently takes one, and I don't know whether smaller tapes exist
for that drive), a dollar or two per tape is the kind of price
differential that could tip the scales whether it's cheaper to go with
two small tapes or one big one. Not down in the noise at all.
/~\ The ASCII der Mouse
\ / Ribbon Campaign
X Against HTML mouse at rodents.montreal.qc.ca
/ \ Email! 7D C8 61 52 5D E7 2D 39 4E F1 31 3E E8 B3 27 4B