NetApp initial attempt to acquire Data Domain shocked most people in the market. Insiders tell me even at highest levels in Data Domain and NetApp engineering, no one had any inkling the boards were inking contracts. That makes a lot of sense with years of blood-war between both companies and EMC. So when EMC finally did find out, the results were awkward and hostile. Never a good combination, but it’s a great way to keep the PR firms busy.
In addition to the dramatic overtures by Joe Tucci in the aggressive courting of Data Domain, with all-cash offers, on top of the cool 100M antes to the original offer, and even bigger cash offers, you start to wonder if this is love or desperation at work? And in the end, Data Domain still holds out for the fiscally less enticing, yet rumored to be contractually binding marriage to NetApp.
What does this signify?
Validation of Data Domain, and the first big move to commoditize the duplication market. Sure, deduplication is a feature applied to primary and secondary disk architectures, but the bigger story is how deduplication is part of the gateway to the next big wave of IT infrastructure. Right now deduplication is fragmented and promoted by vendors all with their own secret sauce. Everyone is special. Everyone is different. In the future, you’ll be buying this feature set from 1 of a handful of market leaders. Same old story for a very disruptive technology feature.
Tucci doth protest too much?
The EMC story for deduplication is fragmented. You have a stalled relationship with FalconSTOR, after mammoth generation-1 VTL deployments from 2003 on, yet no adoption of FalconSTOR deduplication software. You next have Avamar, which is a separate platform specifically engineered for backup. Then you have Centera (which is not performing deduplication, but single-instance store of file objects, with industry rumors the platform is approaching end of development/life. Overall there is a boat-load of non-deduplicated kit spread around the world and no smooth transition for legacy customers.
Then there was about 12 months of EMC telling customers deduplication is ‘in development. Next you have the Quantum partnership, which is in pole position for EMC deduplication / VTL offerings. Finally you have the hostile bid to take over Data Domain. This doesn’t exactly send the good vibes to Quantum, even if you are the CEO .
EMC can spin just about anything out of ether, but my take is they desperately need a new technology base for deduplication and this is much more than picking up market share in the mid-market. EMC knows market share and can market any thing, any where. We all know that. And if you look at other major EMC acquisitions (such as RSA, VMware, etc.), none of these moves were marriages of convenience. EMC is playing strategically and working to fill a technology gap. Even though they have quite a few varations on the deduplication theme, there’s still a gap.
Will the Data Domain & NetApp love last?
You never really know, but my personal theory is this is a mix of practical business sense and high-level hatred of EMC. Quite a few of Data Domain’s original crew came from NetApp. The culture of strong product engineering and R&D are pervasive. Sure, Data Domain reminds me of EMC too , but who can blame a publicly traded company who IPO’s, as we careened into worldwide economic meltdown 2008, for being aggressive. The NetApp deduplication / VTL offering was late to market and in spite of the company reputation for making outstanding and easy to use products, late in this game means you loose. Plus, an injection of aggressive won’t hurt NetApp’s position in the backup infrastructure market.
The most promising aspect is that big company R&D budgeting, strong synergy in engineering disciplines, and the cornering of the NAS primary and secondary storage market will drive out price and flatten the market for deduplicated disk. NetApp already has deduplication in production for NAS filers, now they’re gaining thousands of clients and a proven architecture for backup infrastructure. Plus, this may help Data Domain scale out of 35 usable TB per instance faster than planned.