The new release of NetBackup 5.0 has disk backup capabilities. They call is
a DSSU...Disk <something> Staging Unit. Backups can be directed to the DSSU
and are then de-staged to tape by the media server as time permits. It is
also cache'd on the DSSU, just like CACHE=YES on a storage pool. I don't
know what kinda performance you get from this, but you can have more
simultaneous bakcups running than you have tape drives.

Veritas now also has what they call Synthetic Backups. This is a media
server based process that combines the latest FULL backup with all the
incremental backups (applying deleted and moved files) to create a new FULL
backup without having to actually do a FULL backup. New to 5.0, too.

I'm not endorsing Veritas (I make my living with TSM!)...just adding to the
thread.

Bill Boyer
DSS, Inc.

-----Original Message-----
From: ADSM: Dist Stor Manager [mailto:ADSM-L AT VM.MARIST DOT EDU]On Behalf Of
Stapleton, Mark
Sent: Monday, July 19, 2004 1:21 PM
To: ADSM-L AT VM.MARIST DOT EDU
Subject: Re: Multiplexing to a single tape


From: ADSM: Dist Stor Manager [mailto:ADSM-L AT VM.MARIST DOT EDU] On
Behalf Of Gill, Geoffrey L.
>I was wondering if anyone was multiplexing backups to a single tape.
>Recently I attended a Veritas class, our "other" new backup
>system, (yes I
>did ask to stay completely with TSM) and they really touted
>this as what
>seemed like a very good means of backup. Unless you have the
>newer faster
>technology tape drives I'm not sure I agree with it but don't have
>experience since disk pools have always been plentiful. I'm
>not sure I agree
>with it anyway because I believe restores would be slow since
>data would be
>scattered throughout the tape. What if multiple restores need
>the same tape?
>Can they run at the same time and if so wouldn't that be even worse?

[sound of hollow laughter]

Veritas doesn't tell you about the sting in the tail of their tape
multiplexing. While it can make for faster backups, particularly with
libraries with a small number of drives, they neglect to tell you that
the de-multiplexing necessary to perform restores makes such restores
considerably slower--as in (sometimes) an order of magnitude slower.

...and after all, I personally don't give a hoot about faster backups;
it is faster *restores* I'm after.

--
Mark Stapleton