News Stay informed about the latest enterprise technology news and product updates.

Data delight: A 'set and forget' solution for classifying data?

Following a time-honored principle in computer science, a storage startup has put intelligence into the physical device and introduced what one analyst says could be a game-changing event in how you manage data.

What if you could shuffle data up and down your storage chain without lifting a finger or paying through the nose?

Data classification -- the holy grail of storage -- has claimed its share of victims. Sorting data according to type and matching it to the appropriate tier of storage sounds like a prudent approach, but making it happen costs time and labor. It's hard to find tools to handle the many types of data that are stored; hard to ensure those tools support your platform and hard to actually make those tools work across data centers, or even within one location.

"From a purely software perspective, trying to classify your data in order to properly tier it is actually a pretty daunting task, at least with the state of the art today," said Dave Russell, a research vice president who specializes in storage at Stamford, Conn.-based Gartner Inc.

It's like cleaning your house. You can't do it once and be good for the year.
Dave Russell
When the classifying is done then begins the daunting task of actually moving the data -- and the need for more tools, Russell said. And, once all systems are go, get ready to repeat, because data changes over time. "It's like cleaning your house. You can't do it once and be good for the year," he said.

Compellent Technologies Inc., a four-year old storage provider started by the founders of Xiotech Corp., promises a less labor-intensive approach. Its storage area network (SAN), called the Storage Center, comes with a "data progression" feature that automatically classifies and migrates data to the appropriate tier of storage, based on frequency of use. The assumption is that data that has not been used for a while has less value than data that is accessed frequently.

Most applications write small amounts of new data to a single large file, such as the Outlook.pst.file. Compellent, based in Eden Prairie, Minn., claims to offer the industry's first application to classify and move data between storage tiers at the block level.

Compellent's Data Progression tracks a small amount of metadata on each block of stored data. The metadata, such as frequency of access, is used to allow the data to move up and down among the different tiers of storage. New data is written to the fast Fibre Channel disks and then over time the data is migrated to the slower and much cheaper Serial Advanced Technology Attachment disks.

"It's a different paradigm," said Bruce Kornfeld, vice president of marketing at Compellent who was formerly with Dell. "Our customers' biggest trepidation? Literal disbelief that the system is actually doing what we say it is doing."

Game-changing event

The different paradigm intrigued IWCO Direct, a direct mail house in Chanhassen, Minn. When IWCO moved from a worn-out mainframe platform to Windows 2003 servers a few years ago, it looked at all the big SAN vendors, and settled on Storage Center from fledgling Compellent.

Data management is a big deal at IWCO. The company, which does about $120 million in sales according to Hoover's Inc., turns names and addresses into marketing campaigns for its customers. Starting with a mailing list, IWCO checks the addresses, composes the marketing pitches, produces the customized mail pieces and packs them off to the post office. The company pumps out 250 million pieces of mail a month and is "growing like crazy," said Rodger Smith, enterprise architect at IWCO. Most jobs take about 35 days to get through production.

More on storage

Users get mixed results tiering storage

Taking the tiers out of tiered storage

CIOs fighting uphill battle against renegade apps

"We are turning over between 1.5 and 2 terabytes a month of data online," Smith said. "We need to manage all the current data, and we have to have a way to keep the older data out of production but still available, in case we get audited by the post office or a customer comes back with a question." IWCO wouldn't be in business without its storage component, Smith said. Nor could it afford to keep all its data on primary storage.

Despite some early glitches with the configuration of the Compellent SAN, IWCO jumped at the chance to work with the company on its Data Progression. IWCO also uses CommVault Inc.'s Data Migrator product. Smith estimates the company has conservatively saved more than 5% in capital investment in its storage management hardware and software, even after adding in the cost of the tier management products. It has also saved on labor costs. Smith says the company would need "at least another half full-time employee" to pick up the administration now handled by the automated tools.

Gartner's Russell has been impressed with the offering. "Someone like Compellent stepped back and said, 'Yes information lifecyle management is hard. Wouldn't it make sense to actually do that on the storage device?' So rather than going up the stack on all the different servers, they applied some intelligence down lower on the physical storage device. That's a time-honored principle in computer science," Russell said.

Russell said he finds Compellent's dynamic block architecture compelling on a couple of fronts. "You don't have all these interoperability/matrix issues, you don't have to worry about many layers of management software to try to migrate the data, and maybe the third piece is the most impactful: It's a little bit of a 'set it and forget it mentality,'" he said.

And in the grand scheme of storage? Is the product something nice to have or would he call it revolutionary?

"It's always scary to call something revolutionary. But I think it's closer to that than to a nice-to-have option and it also goes beyond an incremental improvement -- doing what you did before a little cheaper and faster," Russell said. "Yes, I think it is closer to a game-changing event because, No. 1, it works with all kinds of data. That's powerful. And No. 2, it is automatic."

Let us know what you think about the story; e-mail: Linda Tucci, Senior News Writer

Dig Deeper on Enterprise data storage management

Start the conversation

Send me notifications when other members comment.

Please create a username to comment.