NetApp Inc. today made a bold move to become the market leader in data deduplication technology when it acquired Data Domain Inc. for $1.5 billion in a mixed cash and stock transaction.
The deal is expected to close within 60 to 120 days. NetApp paid $25 per share – a high price considering Data Domain's shares opened at $17.40 today – but it acquired the company generally considered the leader in one of the hottest technology markets.
Jay Kidd, NetApp's chief marketing officer, said buying Data Domain would allow NetApp to compete in network-attached storage (NAS)-based disk-based backup, as opposed to the virtual tape library (VTL) space where it offers deduplication with NetApp NearStore VTL. Currently, the virtual tape library is NetApp's only backup hardware product offering with deduplication, and the deduplication it offers with its filers isn't optimized for highly sequential backup workloads, Kidd said.
On a conference call with press and analysts Wednesday night, Kidd emphasized that the NearStore VTL will stick around for those customers who are still interested in using tape, but "Data Domain will help us compete in an increasing number of installs wanting to minimize their reliance on tape," he said.
However, analysts are skeptical about the idea that NearStore will coexist alongside Data Domain's offerings. Data Domain's focus is on disk-as-disk without the VTL interface, but it does offer a virtual tape library appliance. That appliance doesn't integrate with back-end tape, which NearStore does, but NearStore doesn't do replication with dedupe, while Data Domain offers replication with dedupe.
"You could add back-end tape integration to Data Domain pretty easily from what I've seen," said data backup expert W. Curtis Preston. "It wouldn't be as easy to add replication to the NetApp product."
Arun Taneja, founder and consulting analyst at Hopkinton, Mass.-based Taneja Group, also pointed out that Data Domain's variable-length segment dedupe algorithm for backups is generally considered stronger than NetApp's NearStore block-level approach.
"The current [NetApp] dedupe offering hasn't been anything to write home about," he said. "Variable chunks don't always look for duplicates at block boundaries, which means it'll pick up a single-character difference between files."
The price of the deal only amplified skepticism that the Data Domain products will remain relegated to one part of NetApp's business. "You have to believe there will be some product rationalization, where they choose the best [product] for primary [storage] and the best for secondary [storage]," said Brian Babineau, a senior analyst at Milford, Mass.-based Enterprise Strategy Group.
NetApp tried to downplay the overlap between Data Domain's offerings and its own, but both companies position deduplication products with a NAS interface for nearline storage and archival storage, albeit with different approaches to dedupe. NetApp's dedupe for nearline storage is post-process, while Data Domain's operates inline.
NetApp officials on Wednesday night's earnings call said NetApp tends to be strongest in nearline and archiving deals where the primary storage is also a NetApp device because customers can incorporate NetApp's native data replication and migration tools. Kidd said NetApp intends to use Data Domain's nearline product to compete in mixed-vendor primary storage environments.
There's no overlap on dedupe for primary data -- Data Domain has stayed away from that and concentrated on backup and archiving. Analysts are split on whether NetApp will try to use Data Domain technology to enhance its primary dedupe. "I think you could see the [NetApp] dedupe enhanced with the Data Domain algorithm," said Dave Russell, a vice president with Gartner Research. "I just don't see spending that kind of money without really trying to get the value out of dedupe technology across the portfolio."
Countered Preston: "With this acquisition NetApp becomes the first company I'm aware of that has both a solidly inline and a solidly post-process dedupe product, and they are totally different beasts."
NetApp's dedupe works for primary storage because it analyzes only changed blocks in an environment where dedupe ratios are much lower than in the backup world. As a post-process offering, it doesn't sit in the data path or interfere with performance during production hours. Data Domain's product is optimized for high deduplication ratios on highly repetitive data and sits in the data path.
Preston also pointed out that Data Domain's licensing doesn't necessarily make its software a good fit for archiving. "From a pricing perspective, [NetApp deduplication] is probably much stronger for nearline storage," he said. "Data Domain's licensing is based on very high deduplication ratios and might be much more expensive gigabyte-for-gigabyte with a lower ratio."
Meanwhile, according to Taneja, this deal raises the likelihood of further consolidation in the dedupe space. "I believe that at some point in time [NetApp rival] EMC [Corp.] is going to have to buy [deduplication partner] Quantum [Corp.]," he said, "simply because deduplication technology is that important a competitive weapon now."
