Storage Tip: Mitigate pain of storage growth with active archiving, data deletion, and data de-duplication

by David Hill

November 28, 2007 —

What seems to be the problem? Estimates of annual storage growth are 60% and
annual budget growth around 4-5%. While your storage growth may not be that
high, the growth rate of storage is still likely to exceed the annual overall
decline in storage prices. Can you afford it? This problem isn't new, nor is
it only about how to squeeze out enough dollars to pay the bill. It's also a
question of management. Throw in power and cooling requirements, and will you
be able to handle the extra environmental demands?

What you need to know: Unfortunately, the problem will likely continue for
years to come. Not only must you do things better; you must do different things, and that requires a transformation in how you go about doing the storage business.
Note that the pressure of storage growth is only one of the elements driving
that transformation, but it is still enough to get you started thinking about
the problem. (Previous storage tips have already touched upon some of the drivers
and future storage tips will continue to do so.)

Actions that you can take include active archiving, data deletion, and data
de-duplication.

Active archiving can be the repository for your fixed content data. Since this
is likely to be the bulk of your data, what you do with active archiving data
can have a significant impact. Let's see why. First, your active production
data becomes smaller with the burden of fixed content data being co-mingled
within it. Thus, there is less data to be backed up. Active archiving data does
not necessarily have to be backed up in the traditional sense. Rather it must
be replicated -- which is not backup -- and the number of replications is likely
to be less than the number of copies that can be found on backups. Second, active
archive data is typically stored on a different tier of storage than is active
production data. That tier of storage may well be SATA disks that are more cost-efficient
than the FC disks upon which active production data is typically housed. But
you save not only on the disks themselves, but also on power and cooling as
SATA disks are greener than FC disks. (They use about the same power, but store
more data so you can store more data for the same power requirements.) Finally,
active archives can be subject to data de-duplication techniques, such as single
instance storage. Moreover, apart from cost and green savings, an active archive
is essential for data retention management of your data.

And data retention management gets us into the subject of data deletion. Get
rid of data that serves no useful business purpose. Yes, you must take into
account all the legal requirements for when you can safely delete data. But
once that has been done, put in place active policies for data deletion. Some
work may be required to put the necessary data deletion policies in place. Remember
that you need to delete not only the original copy of the data, but all the
copies that were used to "protect" that data wasteland. You must make
sure that data deletion takes place on both active archives and backup copies.

Finally, apply data de-duplication when possible. Now data de-duplication is
a feature of a product and not a product in-and-of itself. But data de-duplication
can be used in disk-based backup as well as in active archiving. Don't protect
unnecessary redundant copies of data.

What can you do about it? What you cannot do is sit idle. Transformation can
be unpleasant as it can require changing processes with which you are comfortable.
Moreover, transformation may have implications for who does what and how tasks
are done and you may not be totally comfortable with the changes. But those
changes may also relieve the pressures that the growth of storage is causing
and make your life easier overall. Lead the charge. The changes are more likely
to be acceptable that way and you are more likely to be recognized as a person
who can get things done right.

Mesabi Group