I won't call this comparing apples to oranges but it might be Granny Smiths
to Red Delicious...
I suggest to all our clients that they run TSM client software compression
and 90% do but some don't (and often it is because of valid reasons such as
servers that house only blah.tar.Z already compressed files, etc...)
BUT I also run with the most compression I can get our of my TSM server's
3590 tape drives...
so a lot of data is being 1st software compressed by the client and then
hardware compressed by my 3590 drives...
I'll list some numbers below but basically it seems like I only get 91+% of
my tape capacity for one environment and 88.7+% of my tape capacity for
another environment with double compression.
All this is based on how much you trust the numbers because in an
environment that I DON'T DOUBLE COMPRESS...
check the numbers at the very bottom ;-)
What do I see...
Estimated Capacity(MB) (full 3590-J tapes in 3590-B1A drives, with no estcap
set)
---------
  9,191.0
  9,191.0
  9,193.6
  9,112.9
  9,952.8
  9,356.7
  9,200.2
  9,220.9
and Estimated Capacity(MB) (full 3590-K tapes in 3590-E1A drives, with no
estcap set)
---------
 35,995.9
 35,995.9
 35,407.7
 35,377.0
 35,549.4
 35,547.0
 35,658.3
 35,328.8
and Estimated Capacity(MB) (full 3590-K tapes in 3590-E1A drives, with no
estcap set)
NO CLIENT COMPRESSION...
---------
253,780.6
253,780.6
216,788.5
250,530.4
258,249.1
788,268.4
181,304.0
213,428.1
241,045.4
256,423.9
249,354.7