In a recent blog Tom Cook, CEO of Permabit declared that all-flash array vendors, which count on deduplication as their key advantage, are going to disappoint their investors or shareholders. His confidence comes from the rapid adoption of Permabit’s deduplication appliance, SANblox, by tier 1 storage vendors. Those tier 1 vendors, he feels now have a level playing field with the start up vendors. While Permabit’s success with tier-1 vendors is impressive, and it certainly helps tier-1 vendors compete with the likes of Pure Storage and Nimble, does SanBlox render many of the flash storage startups impotent?
Will All-Flash Startups Disappoint?
For an all-flash startup to disappoint their shareholders or investors assumes that the only feature they bring to the market is data efficiency. It also assumes a high level of execution from the tier-1 storage vendors. Clearly, data efficiency (deduplication and compression) has proven itself to be a table stakes requirement for flash storage systems. And, if that’s all the system offers then I agree with Mr. Cook, that particular all-flash startup is going to yield disappointing returns. And there certainly are some all-flash array vendors that do fall into a dedupe-only category, so disappointment may be on the horizon.
History Repeats Itself, Again
We’ve seen this play out before. For example when 3PAR, Compellent and EqualLogic came to market they certainly didn’t have deduplication as a feature. They offered thin provisioning, easier volume management, better performance and more aggressive price/performance. Legacy vendors responded by adding those capabilities, clumsily at first, to their offerings, or by buying one of the above companies (and sometimes by doing both). The current round of startups, Pure Storage, SolidFire, Tegile, Nimble, Avere, Data Gravity and others leverage flash to get the attention of prospective buyers and, to Mr. Cook’s point, have counted on deduplication and compression to make the price of a hybrid or all-flash array more attractive.
The next step for the current band of startups is to innovate beyond just selling flash for the same price as high performance disk (which some already have), adding features like scalability, quality of service, high availability, data classification and cloud access. At the same time, these vendors need to keep working on their platforms to make them more flash-ready, reducing latency and increasing performance. But in general the conversation has to move beyond performance, these vendors need to convince perspective IT buyers that they can get more done in less time and for less money with their solutions. When I talk to IT professionals that have made that all-flash system leap of faith, there is a consistent theme. The performance may have gotten their attention, but the ability to massively simplify storage management keeps them loyal.
Another key differentiator for the 2015 round of new storage companies is that much of their executive teams come from one of the aforementioned startups. These executives are good at this, they know what they are doing and they know how to run a business. Their path to success has been well mapped by the successful transitions of 3PAR, Data Domain, EqualLogic and Compellent.
There are some all-flash startup vendors that are selling an off-the-shelf storage server filled with solid state disks (SSD) that’s running open source storage software, with deduplication thrown in. Cook’s prediction about these vendors’ ultimate disappointment may be very accurate, but other all-flash startups are innovating beyond data efficiency.
It is reasonable to assume that a tier 1 vendor that can standardize on a single deduplication engine for their entire storage portfolio would have a tremendous market advantage. Data could be stored efficiently across the enterprise and the amount of data that would actually need to be moved would be minimized. But the tier 1 vendors have to be able to execute this strategy, and many of them have proven to be a disappointment in this area. Most, not all, tier 1 vendor’s storage portfolios are a hodgepodge of products with little synergy and even less vision.
In the end, both tier 1 storage vendors and startups need to keep innovating, solving customer problems through better performance and easing the storage management burden. Either type of company that fails to do this will put its future at risk and disappoint its shareholders or investors.
Finally, innovation needs to be about software, for sure, but also hardware. Hardware needs to evolve so that it can keep pace with memory-based storage. Vendors also need to provide data mobility, so that complete data sets can be moved between storage systems, data centers and even into the cloud.