Network Computing is part of the Informa Tech Division of Informa PLC

This site is operated by a business or businesses owned by Informa PLC and all copyright resides with them. Informa PLC's registered office is 5 Howick Place, London SW1P 1WG. Registered in England and Wales. Number 8860726.

Disk For Archive Is Not Dead

In his recent blog "Copan's MAID Fades away," 3PAR's Marc Farley
predicts that MAID is dead and that we can forget about disk for
archiving applications. I usually agree with Marc but this time not so
much. I believe that MAID is not dead and more importantly,
neither is disk for archiving. MAID (massive array of idle disks) has already moved on. The challenge that Copan Systems had, in my
opinion, was that it was a stand alone system with a single purpose. If
their sales people and engineers couldn't convince you that MAID was
the right solution they had no other option. Comparatively companies
like EMC, Nexsan and Hitachi have implemented spin down technology as
merely a feature within their arrays. You can implement MAID now or in
the future. Its up to you, but it's there when you are ready. Also some
of these companies can offer several stages of MAID that will provide
some power savings at significantly less performance loss as opposed to
the all or nothing condition that Copan presented.

Disk archive is also gaining traction. The economics of a disk
archiving project are too compelling not to consider in this economy.
In disk archiving, spin down is just one potential capability amidst a
host of other requirements. In disk archive you need the ability to
ensure data integrity over the course of a decade or longer, the
ability to scale the archive infinitely, the ability to optimize the
storage of the data with deduplication and compression, the ability to
find data within the archive and potentially the ability to maintain a
chain of custody with WORM lock down of files.

Potentially, most important is how moving data to the disk archive has
evolved. Most disk archive systems have moved from a proprietary API
access to a network file system access like CIFS or NFS. While APIs
have a role to play, having the simple access of a network mount point
makes adoption easier. It allows customers to manually move data to the
archive, as we discussed in our blog "Manual Moves" over at Information
Week or to leverage File Virtualization techniques, as we discussed in
our recent article "What is File Virtualization?".

To some extent I think Copan fell victim to the first mover
disadvantage
, they were very early to market with MAID and dense
storage. Since their first systems, which were primarily focused on
being a disk backup solution, much has changed. Power management was
not the concern for as broad a section of customers as it is
today. They had to take most of the arrows for explaining what MAID
was, and they had the sole responsibility to convince people that drives
would power back on when you needed them to. Remember disk aerobics?
Now MAID is much less of a all or nothing decision. You can ease into
it as you feel it makes sense for your environment.

On the archive side, cost of disk continues to plummet, moving data to
the archive has become significantly easier, but most importantly, the
need is more severe. When Copan first opened the doors unstructured
data was not the critical problem that it is today. File based data is
drowning many data centers, and I think we have just now reached the
tipping point where storage administrators are prepared to make
significant investments in unstructured data management processes.