April 06, 2009, 4:46 PM — Tasty Baking Co. produces more than 4.8 million cakes, doughnuts, cookies and pies each day. And the Philadelphia-based snack food giant also manages to generate another crucial commodity -- computer data -- in equally impressive amounts. "It's something to be concerned about," says Brendan O'Malley, Tasty Baking's vice president and CIO.
Like a growing number of businesses worldwide, Tasty Baking is facing a data explosion. The use of e-mail and rich media applications plus the need to stay on top of regulatory compliance are stretching storage resources to the limit at a time when budgets are shrinking. Handling rapidly spiraling storage needs without spending tons of money is a challenge that's facing just about all IT managers. O'Malley crystallizes the need into a single phrase: "More space for the money."
To get a handle on storage demands, IT managers need to carefully examine internal storage practices, use specialized software tools and find appropriate storage systems for various kinds of data. Andrew Reichman, an analyst at Forrester Research Inc., says that they must look first at the two best ways of cutting storage costs. "One is using more-dense drives, and the other is reducing your footprint," he says.
Footprint reduction is relatively easy, says Reichman, asking, "Can you use the Delete key?" But David Hill, founder of The Mesabi Group, a storage management consultancy in Westwood, Mass., notes that many businesses, particularly those in regulated industries, are scared stiff of deleting anything other than spam. "There are a lot of legal reasons, tax reasons," he says. "You have to be careful about how you go about getting rid of data."
Fixed vs. Fluid
Although managers may be reluctant to flush files, they can bring greater efficiencyand cost savings to their storage processes by organizing data more efficiently. Hill recommends separating fixed content -- e-mails, boilerplate documents and other rarely modified files -- from active archives that contain more fluid types of data, such as transactional data.
Virtual Storage Headaches
What is the top challenge related to storage for your virtual server environment?
* Maintaining high performance: 23%
* Completing backups on time: 19%
* Efficiently managing storage capacity: 14%
* Controlling costs: 11%
* Provisioning storage: 8%
* Deploying tiered storage solutions: 7%
* Performing effective application restores: 6%
* Generating accurate and useful reports on the environment: 6%
* Maintaining a high backup success rate: 3%
* Deploying effective multipath solutions: 2%
* Other: 1%
Source:Forrester Research Inc. survey of 124 global IT decision-makers currently using x86 server virtualization technology, September-October, 2008
Hill notes that many businesses unthinkingly use costly, high-performance Fibre Channel storage environments for both types of data. Yet splitting up fixed and active data makes it possible to use less-expensive storage technologies for less volatile information. "You can use what's called capacity storage instead of performance storage," Hill says.
Capacity storage usually comes in the form of Serial ATA drives, which have higher capacity but cost less than other options. "So you get more bang for your buck," Hill says. "That's one thing that businesses can do: move some of their data to these lower-cost drives."
O'Malley agrees that organizing content is a relatively easy way to cut costs. "We have some high-performance applications that run on Fibre Channel, and other applications that don't need that kind of performance that we run on regular SATA types of disks, where we have a better price for performance," he says.
Data de-duplication -- the process of eliminating redundant files -- can also save precious storage space. Backup products that de-duplicate data at the host level include Asigra's Televaulting, EMC's Avamar and Symantec's Veritas NetBackup PureDisk. Disk libraries from Data Domain, Diligent Technologies, EMC, Quantum and Sepaton de-duplicate data at the target level, such as the desktop. Both methods determine whether data segments or files are identical by running a file-level comparison or using a hashing algorithm.
Even relatively insignificant files can pile up to create a big storage headache, says Dave Richerson, manager of infrastructure services at Republic Bank in Louisville, Ky. "If we send a Word document out, that doesn't make a huge impact on the e-mail system," he says. "But 15 people might store that Word document on their home directory as something to refer to down the road and, well, that's 15 copies."
To create space-efficient backups without endangering critical data, Republic uses Avamar for its branch office backups and EMC Disk Library for data center backups. "We've had this system in place for about four months and have seen an immediate, positive improvement in storage efficiency," Richerson says.
De-duplication vendors often claim that their tools can provide a specific ratio of data reduction. Actual data de-duplication ratios can vary widely from customer to customer, of course, but Richerson says he was "floored" by the ratio his deployment achieved. "We see a de-dupe of 8-to-1, or basically 24TB of data is de-duped down to 3TB," he says.
At the cutting edge of storage, virtualization promises to help businesses archive data with unprecedented efficiency. Greg Thompson, director of IT at Northeast Bank, is relying on the EqualLogic PS5000 Series virtualized storage-area network from Dell Inc. to store corporate data. The systems are based on a virtualized modular storage architecture that enables Lewiston, Maine-based Northeast to purchase storage as it is needed, preventing overprovisioning.
For Thompson, the new technology marked a welcome departure from local server-based storage. "Local storage on servers gets unruly," he says. "When you run out of space, what do you do, buy another server? So we decided central storage was the answer for us."
Thompson says he opted to tie central storage together with virtualization for server and storage consolidation. "The whole process has worked really well for us," he says, noting that storage can now be scaled easily and efficiently whenever the need arises.
Thompson also appreciates the virtualized environment's inherent redundancy. "If we do have a critical, fatal error happening here at the primary data center, we can have access to just about all our network data within a matter of hours," he says.
Like Thompson, O'Malley says that virtualization can give a big boost to businesses striving for better storage efficiency. Tasty Baking now uses a FAS3020 system from NetApp Inc. (download PDF) to provide Fibre Channel-based storage resources to 40 virtual servers running on 10 physical servers. "We moved away from having any local storage on servers or having any sort of direct-attached storage or anything of that kind," O'Malley says. "Now everything runs off NetApp."
Before O'Malley and his team moved to the virtualized environment, adding storage was both a challenge and a puzzle. "If we were out of storage on a particular server, we wondered if we should find more slots where we could put more disks in, or maybe attach some storage directly to that box," he recalls. "You might have storage space on another server, but you couldn't really share it over to that machine."
The new environment allows O'Malley to consolidate and optimize storage resources. "We've got one place where everybody is going for storage," he says. "With the thin provisioning, and the ability to move storage, we have one bucket that we're pulling from, and we can get by with a lot less overhead."
The virtualized arrangement allows for more planning flexibility. "We don't have to determine what the maximum possible storage might be for an application over the next three years," O'Malley says. "We don't have to ask that question before we start and advise the server to hold that."
Virtualization also frees storage from constraints imposed by a server's physical configuration. "It's now just a question of finding more storage from the single pool on NetApp," O'Malley says.
He notes that storage efficiency and simplicity meshes well with Tasty Baking's corporate philosophy. "When you think about it, we're just like a corner bakery, only a thousand times bigger," O'Malley says. "What's key for us is keeping our environments as simple as we can."
Edwards is a freelance writer in Gilbert, Ariz. Contact him at firstname.lastname@example.org.