Download the authoritative guide: Enterprise Data Storage 2018: Optimizing Your Storage Infrastructure
When I started to write this article, I expected to launch immediately into innovative products and vendors because innovation, like adventure, is out there. But the deeper I got, the more I realized that innovation is very much in the eye of the beholder, and I was going to have to define my terms and assumptions.
Any storage company that is still in business is trying to innovate. If innovation means a new and improved approach to a problem, then very few storage vendors are stuck in place just waiting for the market to pass them by.
Sure, the start-ups define themselves as innovators – some may even be innovative. They are looking to create a market and ride that momentum. But the established vendors are busy innovating too, and they have a customer base to sell to that the startups do not.
And just because you’re a startup with a hot technology and funding, doesn’t mean you’re innovative – you’re not really much of anything until you sell enough product to keep yourself in business — or sell your company to the biggest bidder. As blogger Justin Parisi put it, "There are definitely some real innovative products out there and some real good talent working at these companies. I have a lot of friends and ex co-workers at startups. But there is also a lot of redundancy disguised as innovation."
And really, what’s innovative?
A giant disruptive technology that does not sell well at first but fundamentally changes a market is innovative. Flash comes to mind, which took years to find its present level of acceptance in the data center – and even now is installed at less than 10 percent of enterprise data centers.
Most companies try to avoid building a business on disruptive technologies no matter how innovative they might be, because positive commercial impact is measured in years. Their income model may be to hang on until they achieve market acceptance, but their VCs usually have other ideas in mind.
Established vendors can afford to pour in a certain amount of R&D to innovation, or they can afford to buy start-ups with innovative new technologies that actually work. EMC did this in 2014 by acquiring DSSD flash technology, then Dell acquired EMC this year and thus DSSD. Apparently Michael Dell is pretty excited about it.
Also, remember that all storage innovations must answer similar data center dilemmas: Admins need to grow their storage capacity in response to fast-growing data. They need to store data safely and they need their users to be able to find it again. If the data is lost or corrupted, they need to be able to restore it according to application service levels. As servers get faster and applications bigger, they need to accelerate storage performance. They need to validate that their data protection and disaster recovery processes are working. They need to avoid forklift replacements and support legacy systems (eventually every system becomes legacy). And they need to do all this in budget.
So Is Innovation Real?
Remember Apple’s famous Macintosh commercial from 1984? The personal computer already existed, but Steve Jobs was able to sell an innovative new interface. And in so doing, he redefined the personal computer market.
Although Apple is a story so big that a major studio made a movie out of it, innovation continues today. Here are the guidelines I kept in mind for selecting storage innovation to highlight:
- I deliberately left out mind bending R&D. This included projects like helium drives, DNA molecules (yes, DNA as in human), holograms (Star Trek anyone?) and the use of quantum physics for instantaneous data copies. They’re based on real science and may even come to market someday. But right now they’re nowhere near ready to productize commercially.
- Speaking of which, it’s an actual selling product. I don’t mean to slam storage start-ups. Our biggest and most successful companies all started out as start-ups, and I know several that are brilliant and will no doubt succeed. The issue is that any start-up can claim that it’s developing new storage that is so great it’s about to make a moon shot (in which case, someone call Richard Branson). Start-ups do not have dibs on innovation and must prove themselves in the market like any other storage company.
- Innovation exists at established and new vendors. The press tends to align innovation with start-ups, and the established vendors do not get nearly as much attention on the innovation side. Less attention may be justified when the vendor is spending large amounts of time and money just to keep a legacy product in business. But often it’s the established vendors who have the money and will to launch innovative research into new technologies.
- As I’m not writing a tome, I must pick and choose innovations. Many vendors and VCs are pouring money into various technologies that they hope will innovate the marketplace. Leaving out the majority of them says nothing about whether I think they’re innovative or not, it just means that of necessity I must keep a narrow focus.
So rather than calling out various innovative vendors, I delved into three large storage segments that have big challenges and need big solutions: data awareness in primary storage, storage IO optimization and archiving. These are by no means the only innovative segments out there; innovation abounds in hyperconvergence, data protection, data reduction, cloud storage/failover, and more. But one must start somewhere.
- Data awareness in primary storage. These production arrays are purpose-built for detailed metadata capture and reporting, analytics and additional processes around search and governance. Qumulo and DataGravity are doing some very interesting work here. Honorable mention will go to software products like Acaveo that sniff out dark data, including primary data, from multiple silos.
- Storage IO optimization. Most high-performance arrays have introduced hybrid or all-flash models. Flash, of course, speeds up storage performance, but the innovation comes in QoS features like workload optimization. QLogic is a good example of using decades-old Fibre Channel for an innovative approach to array-based performance. However, not all IO optimization comes from the storage side. In future articles, I’ll also discuss what I consider the ideal place for optimizing IO: the server. Condusiv Technologies, Infinio, VMware and PernixData’s FVP are in this camp.
- Archiving advances. Yes, archiving. Archiving seems like the least likely place for innovation, but the reverse is true. Two realities of the storage market are driving archiving innovation: huge data growth and the push to wring more value out of data. Storiant, SpectraLogic, and Fujifilm Dternity are doing some very interesting work in deep archives, analytics and governance.
Photo courtesy of Shutterstock.