So far, most of the action in storage has been towards
pulling storage resources out of individual machines and placing them on the
network. The move has been from direct attached storage to network attached one,
and eventually to a completely separate storage area network. On one side, this
evolutionary process has helped consolidate storage resources, and given
organizations the much-needed space for their growing data. On the other side,
it's also opened up a Pandora's box of new challenges. For instance, now
there are so many different types of storage devices that it's become a
challenge to get them to talk to one another seamlessly.
In This Story |
Storage Virtualization |
What's needed now is a set of
technologies that can manage them; something that sees all
storage resources as a single pool, and allocates it automatically to
applications and servers as the need arises. Not only that, but now you also
need technologies that can seamlessly add more storage to this pool as and when
needed, and it has to be as easy as joining a few Lego blocks. As data volumes
increase, the storage pool shouldn't slow down by the inertia of its own mass.
So interfaces between different resources have to be designed such that they can
keep the data flow consistent. Therefore, lots of work is required on the
interface technologies for storage networks. They must be able to converse in
the same language.
Besides this, work also has to happen on the data management
front. Technologies are needed that can differentiate data based on its
relevance, and move out what's not needed to slower storage devices. It has to
be intelligent enough to determine when the data would need to be retrieved
again, if at all. This has to happen fast enough else you'll end up having one
big blob of useless data.
Data growth also poses another
challenge, of security. For this you need technologies that
can differentiate data based on its criticality. Highly critical data needs a
more secure storage space than moderately critical one. You need technologies
that can control access to different storage pools. All this may sound
impossible, but the technologies that we're going to talk about in the
following pages aim to address all these objectives. Storage virtualization for
instance aims to convert the storage resources into an easy to manage pool,
whereas SAS or Serial Attached SCSI aims to get all the interconnects right.
Disk based backup devices aim to improve the speed of data backup so that the
storage pool always remains agile. Likewise, e-mail archiving solutions are
trying to manage the ever-growing volume of e-mails in organizations. As e-mail
is being increasingly used as a business communication tool, you can't afford
to lose any of it. At the same time, you can't let it lie on every
individual's PC or even on the mail server. Hence, e-mail archiving is the
solution.
Similarly, there are technologies that aim to pack more
storage capacity in lesser space. This is happening on existing hard drives, as
well as newer technologies like holographic and optical storage. Security is
another concern. Now that storage is present directly on the network, it needs
more stringent security measures.
So if you think that all the action is over in the storage
world, think again. It's far from over. In fact, it's just begun.
Anil Chopra, Rinku Tyagi, Sujay V Sarma and Swapnil Arora