>I'm assuming that the MASTER LIST of these 1000 files, their creation, modifiy
>date, # of generations kept, how many there currently are, etc. are kept in the
>TSM Server Database.   If we watch an incremental backup, via the GUI, we are
>seeing that for this client the COMPARE time is horrible.  It takes FOREVER to
>get to the file to be backed up.

Do you say this on the basis of having run a client trace to gather
statistics, or is this a subjective judgement?  What you may be seeing is the
performance of *anything* going through that file system, which you can gauge
with operating system file traversal commands, like the Unix 'find' command.
Any file system backup facility has to traverse the file system, and depending
upon the topology, disk speed, host load, and number of files in the
directories, can take considerable time.  This is not a TSM problem.
You could also do a comparative -incrbydate backup, which does not gather
information from the server and only checks files for newness.  (In a normal
incremental, the server is called upon to amass its list of filespace files
and send them to the client for sorting and comparison.)

Perform tests in your operating system to gather baseline figures on your file
systems so as to gauge the time that TSM is taking to back it up.

   Richard Sims, BU
 .