As some of you may know, KeyLabs is not the most reputable testing
organization.  KeyLabs was the firm that reportedly tested Windows
NT against Linux a year or so ago.  They're results showed that NT
was faster in every category.  When the report was published,
members of the Linux community were in an uproar because they had
not been invited to the test.  As it turns out Microsoft funded the
performance evaluation, and had SEs out at the KeyLabs facility to
guarantee the best performance out of NT, while a Linux person was
nowhere to be seen.

As the first paragraph states "performance testing was completed at
the request of Veritas."  I would wage money there was no IBM or
Tivoli involvement with the entire process.


--
Joshua S. Bassi
Joshua S. Bassi
IBM Certified - AIX Support, HACMP,
Enterprise Disk(Shark)& Tape Solutions
Tivoli Certified Consultant - ADSM/TSM
jbassi AT gloryworks DOT com

"Klatt, Michael" wrote:

> http://www.keylabs.com/results/veritas/veritas.html

Intersting...  The test doesn't seem to test real world scenarios.  I wonder
why they didn't test network backup and restore.  I wonder why they tested
client and server on the same machine - how many people do that?  Only one
test
went to a disk cache, the rest were directly to tape.  Did they use client
side
compression on TSM?  That would affect CPU and througput - especially on the
same machine.

I would like to see a test of the products in a scenario that matches how
the
products are used on a daily basis over time.  For example, after measuring
backup times each day for a month in a controlled environment, how do they
compare?


--
Ray Schafer       |    schafer AT tkg DOT com     |    http://www.tkg.com
Ray Schafer       |    schafer AT tkg DOT com     |    http://www.tkg.com
The Kernel Group <=-=-=-=-=-=-=-=-=-> Core Confidence for E-Business
 TKG's Bare Metal Restore can re-build your machine from the ground
   up with a single command - using the data stored in TSM (ADSM)