> Interesting... That hasn't been my experience at all. I recently
migrated
> from ADSM/VM to TSM on zLinux and didn't see a significant difference
in
> MIPs consumption. 

It also seems to be related to the number (and complexity) of the
clients. There is a floor value of resource consumption with TSM that's
unavoidable; the resource consumption increases sharply as you get more
complex devices or start ramping up the number and complexity of
clients. The difference between the straight backup/archive clients and
the application-specific ones also plays a big role (the Exchange and
Notes clients are particularly piggy both on the server and the client
side; in all fairness, they have to do a lot more than the B/A client). 

The CMS version scaled much more linearly; there was an increase as
things ramped up, but it was much slower, and the difference between the
B/A and app clients was much smaller. Supporting a 12,000 client install
on the VM version was doable. No chance with the Linux version. 

In the two sites we deal with who use TSM on Linux, it runs on average
about 250% of the resource consumption of a comparable size Bacula
installation, and cost them about 300% more than the support for Bacula
(if we include the requirement for new tape libraries, etc, the cost is
even higher). I'm not sure why those two sites do it...

If I ever get enough time to finish the CMS Bacula client... *sigh*.
Someday. 

-- db

Reply via email to