I agree with Richard and Dwight.  It depends.  We have client compression
on, I did a bit of testing, and sending client-compressed data on through
the tape drive compression generally doesn't hurt us, but doesn't help much,
either.  In some simple tests I ran, we got at most an additional 10%
compression on 3490 tape drives.

The problem with TSM and figuring out what compression is doing, is that TSM
only tells you what the CLIENT reports sending to it.  It doesn't KNOW what
the hardware compression is doing.  You can't necessarily rely on the
CAPACITY figures it reports for each volume.

Assume, for the sake of simplicity, that the compression ratio is 2:1 for
either client software compression, or your tape hardware compression.  And
assume your "native", or raw physical tape cartridge capacity is 20 GB.

On a client that has 40 GB of data to send, if compression is ON at the
client, it will compress the data down to 20 GB, report 20 GB sent to the
server, and the server will report to YOU that it sent 20 GB to the tape,
and you will have 1 full tape.  That tape volume will show "est capacity" at
20 GB.

On a client that has 40 GB of data to send, if compression is OFF at the
client, it will report 40 GB sent to the server, and the server will tell
you that it sent 40 GB to the tape, and you will still have exactly 1 full
tape, since the hardware will compress the 40 GB down to 20 GB.  That tape
volume will show "est capacity" at 40 GB.

If you have a mixture of clients compressing/ not compressing, you can't
look at the "capacity" figures for your tape  volumes and tell a darn thing.
All you can do is make some controlled tests where you work with a specific
client to send a specific set of data, and see how much data you can send to
a tape before it fills up.  Then you can assume you will get the same
compression ratios on clients with similar data.