We have been using a VTL for 3-4 months and we consider it a success.  

Environment:
------------------------------------------------------------------------
--------
*                       OS: AIX 4330-11
*                      TSM: Version 4, Release 2, Level 1.7
*         Physical Library: Single frame 3494 w/2 3590E1A Drives
*                      VTL: Sepaton ES2100 10.5TB (Raw)
*         VTL Drive Config: 50 Drives
*        VTL Tape Vol Size: 10 GB (picking a size is a religious belief)
* VTL Tape Volume Quantity: 839
*   Avg Nightly Backup Amt: ~750 GB
*             Client Types: AIX, Windoze(NT, 2000, 2003), Oracle (TDP), 
.                           Microsoft SQL (TDP), Novell
*              Collocation: None
*    Total Primary Stg Amt: ~10.3 TB

Problems Leading Requiring Upgrade/VTL Purchase:
------------------------------------------------------------------------
--------
* Poor restoration speed for file servers due to:
- No use of collocation because of single frame 3494, amt of primary
stg, desire
. to make tape management simple (read desire tape library to house all
primary
. tape volumes, no overflow to a cabinet/vault)
- Only 2 drives
- Both drives spinning tapes 95%-99% of day doing migrations, backups
and 
. reclamations, leaving little time for restorations.

* Poor offsite reclamation performance: Offsite tape count growing to
~500 with
. >400 being >50% reclaimable

* Onsite tape count growth forcing expansion of existing
library/purchase of new
. library and associated disk stgpool

Reasons Pushing to VTL over DASD and Physical Library:
------------------------------------------------------------------------
--------
* Ability to do LAN Free back-ups w/o Sanergy
* Storage compression w/o load on TSM server or client
* Performance: Mounts/Unmounts, Reads, Writes (Comparison with physical
tapes
. need to keep in mind your clients/server may not be able to
continuously
. stream data to tape.
* Desire to break out of the structure where growth requires growth of
library
. system AND disk stgpool
* Ease of implementation

Benefits Realized:
------------------------------------------------------------------------
--------
* All of the problems requiring the upgrade have been eliminated
* All primary stg pool data has been moved to the VTL
* 3494 is only used for the offsite COPYPOOL stgpools and offsite DB
back-ups
* offsite COPYPOOL reduced to ~195 tapes with ~5 being >50% reclaimable
* 3494 drives no longer have to spin 95% of the day
* Back-up and restorations greatly improved w/o use of collocation
* Implementation was easy:
- VTL was TSM certified just like a physical library
- VTL install and config: ~1 hours
- TSM Config (Library, Drives, stgpool): ~20 minutes
- TSM labeling of 839 volumes: ~2 hours
- ~2 minutes to update the NEXT STGPOOL on the primary disk stgpools to
point to
. VTL stgpool

We viewed the costs of the VTL vs. a physical library with drives (that
performs like the VTL) and associated disk stgpool to be comparable.

Needless to say, I disagree with the statement that TSM doesn't appear
to be a good fit for a VTL.  Remember a VTL is just a library that can
mount/unmount a tape in less than one second and read/write to the tape
at disk speeds.  Why wouldn't TSM work well with a library on steroids?
 
Thanks,
H. Milton Johnson

 
-----Original Message-----
From: ADSM: Dist Stor Manager [mailto:ADSM-L AT VM.MARIST DOT EDU] On Behalf Of
David E Ehresman
Sent: Monday, November 29, 2004 2:42 PM
To: ADSM-L AT VM.MARIST DOT EDU
Subject: Re: size of active vs. inactive?

>>It has also been discussed several times that TSM doesn't appear to be
a
good fit for a VTL. May want to search the archives to find out more on
the subject.<<

Has anyone heard of good experiences with TSM on a virtual tape library?
Can this ever be a good thing?