The rumor mill is adamant that storage giant EMC is in serious talks to buy Israeli flash-storage startup XtremIO, a move that could trigger an avalanche of flash acquisitions rivaling the scale-out-file-system feeding frenzy a couple of years ago.
Strategically, the move makes sense for EMC. For one, it has promised an all-flash storage array codenamed “Project Thunder,” and XtremIO is building just such a system right now. Talk about good timing. But there’s more, namely the major focus EMC (which is the majority shareholder in VMware) has put on virtualization and big data going forward.
An all-flash array means faster performance across both virtualized and big data environments. Combined with EMC’s server-side PCI flash product called Project Lightning, which keeps hot data in an SSD cache sitting alongside the processor, that’s one powerful hardware platform for tomorrow’s applications.
Keeping up with EMC means going flash
If EMC does pull the trigger, it wouldn’t be surprising to see other flash dominoes fall in a hurry. NetApp, which is allegedly trying to steal XtremIO out from under EMC, would almost certainly have to answer with an acquisition of its own. HP, Dell, IBM, Hitachi Data Systems — everyone, really — would have to kick their flash plans into gear, too.
The reason is this: flash technology is getting cheaper by the day, and new lower prices points — especially when presented in terms of price/performance — are letting flash-storage startups score a lot of customer wins. But flash for primary storage is kind of like open source software, in that until large vendors pin their futures to a technology, it’s difficult for truly big IT buyers to justify investing too much in it. If that technology goes away tomorrow, the CIO who bought it is out of a job.
EMC dropping half a billion on XtremIO and starting to sell all-flash storage arrays makes justifying a flash purchase a lot easier. But if EMC is the only game in town selling flash arrays, then EMC is the only major storage vendor making money off it. And that can’t happen.
Who’ll get bought
Who’s likely on storage heavyweights’ shopping lists if the EMC-XtremIO deal goes through? I’d say Violin Memory is on the top of that list, with Fusion-io not too far behind. They do different things — Violin is flash arrays whereas Fusion-io does primarily flash cache — but they have both have the mindshare and the tech to pay dividends. Other options — which might come at a lower cost than the highly valued Violin and Fusion-io — are Kaminario, which sells some serious systems chock full of flash and DRAM, and Virident, which competes with Fusion-io on the cache front.
After that, a slew of startups that includes Pure Storage, Nimbus Data Systems, Tintri and Nimble Storage have to look appealing. All of these companies have accumulated impressive customer wins and piles of investor cash, and it’s because they can deliver price-performance gains that, presently, most major storage vendors cannot.