NetApp De-Dupes

In a long-expected move, Network Appliance has added block-level de-duplication to its FAS and NearStore R200 product lines. The vendor's Advanced Single Instance Storage (A-SIS) is now offered as a free upgrade to Ontap 7G customers with maintenance contracts; others will pay a "nominal" fee of roughly 1 to 2 percent of the overall cost of their system.

Up to now, NetApp offered A-SIS only in systems bundled with Symantec NetBackup as part of a joint solution.

NetApp's NearStore VTL, which runs under a different operating system, will get its own de-duplication software in mid-June. That announcement is scheduled for the Symantec Vision event in Las Vegas.

The Ontap GX operating system is earmarked to get de-duplication too, but no date's been set. NetApp believes the help with de-dupe is needed primarily in mainstream enterprise applications that use Ontap 7G.

NetApp has tarried in delivering this feature, for which the vendor claims over 50 early adopters. Competitors include Asigra, Quantum, Sepaton, and Symantec, which also support de-duplication of data after it's written to a target disk. Data Domain, Diligent, and EMC (via its Avamar acquisition) de-dupe prior to backup. (See De-Dupe Deep Dive.)Still, NetApp is emphasizing that its target-based de-duplication is different from others in that it will run not only in archiving or nearline storage but in primary storage as well.

"We are making de-duplication available across all applications. It's not limited to just backup environments," says Ravi Thota, director, product marketing, Data Protection and Retention at NetApp. The de-duplication is delivered with RAID 6 capabilities and works with Ontap management features.

This echoes what Jay Kidd, SVP and GM of NetApp's Emerging Products Group, told Byte and Switch last month: "De-dupe is not just useful in backup," he said. "De-dupe in primary storage will get there." (See NetApp's Kidd Talks Turkey.)

At least one analyst thinks this approach distinguishes NetApp from other storage suppliers. "It's interesting that NetApp is the first storage vendor to say you should bring de-dupe into production workloads," says analyst Stephanie Balaouras of Forrester Research. She also compares NetApp's OS-based de-duping to that of Microsoft, with its Single Instance Storage (SIS) feature for Windows Storage Server. (See Analysis: Data De-Duping.)

Despite the claim to be primary-ready, NetApp acknowledges most adoption of de-dupe will be in backup and archiving, with virtual servers and user data or home directories following. Further, NetApp stresses that de-dupe won't be used alone, but with other data reduction features, such as its FlexShare and FlexVol.At least one NetApp partner thinks broader A-SIS support by NetApp will get a good reception. "Customers think it's overdue," says Dave West, VP of marketing and business development at CommVault. He says CommVault implemented hooks for de-duplication by storage vendors over a year ago and supports products from Data Domain and EMC. "We're excited about it," he says. "We think there's a real challenge with managing so many copies in an efficient manner."

The advantages users derive from A-SIS will vary with setup. One adopter, Intuitive Surgical Inc., which makes medical equipment, has realized a 38 percent reduction in space required for its NearStore system archives of primary storage, NetApp says.

Another unnamed customer with VMware reported 88 percent space savings while backing up NFS-based VMware images associated with a mission-critical application from a 64-Tbyte NearStore system to two replication sites.

Traditionally, the problem with de-duping, particularly in software, has been the drain on performance. But NetApp claims the new software has "no perceivable" performance impact; de-dupe consumes about 1 percent overhead in writes and 1 percent to 3 percent in capacity, the vendor says. Still, those figures are internal only.

Mary Jander, Site Editor, Byte and Switch