ADSM-L

Re: Collocation and Disaster Recovery

1998-06-15 10:21:49
Subject: Re: Collocation and Disaster Recovery
From: "Kelly J. Lipp" <lipp AT STORSOL DOT COM>
Date: Mon, 15 Jun 1998 08:21:49 -0600
The point about being selective is this: not all data is created equal.  I
think that most sites have a few nodes that are absolutely critical.  Let's
define critical a little better: applications without whose existence a
business will cease to exist.  Usually this is a pretty small list.  Those
apps must come back first and quickly.  Focus your disaster recovery on
those.  If you look at most disaster recovery plans (when they exist at
all), there is an over-emphasis, IMHO, on data recovery and not nearly
enough on recovering the folks to actually run the recovered data.  By the
time we've got the folks available to run the apps, we can usually have
them all back no matter how long that takes.

Take an example of a 600 GB environment.  Ask reasonable human beings what
percentage of that data is really critical and number is probably something
like 10%.  60 GB is not that hard to recover, if you can find that most
critical 60 GB.  Now obviously some folks have large databases or something
and those folks will rain all over my parade, but most are not in that
situation.

Find you most critical data, ignore Aunt Betty's cookie recipe, restore it
and get on with life.  The real good news about disasters is they don't
happen very often.  Perhaps optimize for things that do happen often, such
as the small disasters: loss of a single file, or small number of files.

Kelly Lipp
Storage Solutions Specialists, Inc.
lipp AT storsol DOT com
www.storsol.com
(719) 531-5926