It looks as if solid-state cache specialist Fusion-io’s IPO on Thursday has investors thinking big about the possibilities of flash storage in the enterprise. Flash array maker Violin Memory raised a $40 million Series C round from strategic public-market investors — on the heels of a $35 million round in February — and a software company called VeloBit emerged from stealth with an undisclosed amount of Series A funding from Fairhaven Capital and Longworth Venture Partners.
In the case of Violin Memory, the new money represents only a fraction of a capital infusion intended to ramp up operations at this critical juncture. CEO Don Basile told me his company also took on $140 million in debt financing, bringing its total amount of new cash to $180 million. I highlighted Violin Memory’s unique position in the flash world — that of an array producer rather than a component manufacturer — a couple of months ago, and it has only gotten stronger. Basile said the market validation of Oracle CEO Larry Ellison’s decision to power his high-powered transaction-processing appliance with flash has been bolstered by EMC’s decision to soup up a number of its arrays, too, with flash as the primary storage layer.
The Oracle situation led to a strong partnership between Violin and HP around creating a competitive appliance, although the EMC validation addresses an even broader market. Regardless, Basile explained, the involvement of large systems vendors in the flash game means that price is coming down fast, which means that companies like his can begin touting flash’s price-performance edge over disk with more gusto than ever. If it’s cheap enough and if customers know that their blue-chip providers such as Oracle, HP and EMC are invested in it, Basile said they’ll very seriously consider buying when their system refresh cycles come around. He says Violin can deliver a far lower price per transaction than can spinning disk.
One issue right now might be the need to write applications to take full advantage of the performance gains of flash, which explains the emergence of VeloBit. Although details are sparse, the company claims to be developing a software product that “allows customers to deploy SSD without changing their applications or primary storage, thus preserving their existing investments in data protection and storage management.” It’s comparable to the situation with multi-core processors, in which hardware advancements outpace advancements in software development.
VeloBit, though, appears to be targeting caching use cases — similar to that which Fusion-io addresses with its components —whereas Violin wants to store companies’ primary data in its arrays.
If flash storage catches on and drops in price like Basile is convinced it’s ready to do, the possibilities are endless in terms of data center evolution. Big data applications are all the rage right now, and flash-based systems would take performance to a new level. Likewise, the notion of delivering flash performance as a service could revolutionize cloud computing and democratize access to Google- or Facebook-like infrastructure. We’ll discuss the future of next-generation distributed systems and cloud storage at Structure 2011 in two weeks, and flash is almost certain to be a big part of that discussion.