What would be the consequential differences between using migration
versus using 'move data' to transfer backup files from one primary
tape pool to another? I see that with 'move data' I would specify
volumes names myself, whereas migration would pick the volumes to
move. Are there other differences that would favor one technique over
the other? A difference in throughput rate would be especially
interesting.

The files are moving between data centers, and are now in virtual
volumes. The target tape pool consists of physical 3592 cartridges.

This is all running on RHEL 4, TSM 5.4.0.3, over a 10 Gb network link.

Thank you,
Keith Arbogast
Indiana University