Historically, the storage industry, simply put, sucks at agreeing on – and deploying – open standards for anything. This fact makes huge sense when you consider that the “standardized” segments of the storage business (e.g., raw disk storage) have survived for years on razor thin margins per disk, while the software and system value-add that have floated on top of this core architecture have been priced at anywhere from rational margins to excess profits. Nobody wants to give up those margins! Startups need those margins to innovate while the major system and storage houses who exert a level of market control simply love the ROI. Nobody really wants an open standard – unless by some chance it is constructed to allow customers to move off the competitor’s offerings and onto “mine.” This win-lose mentality results in a lot of talk (and meetings) about open standards and products, but very little action. SNIA is a tissue paper tiger.
Enter – the cloud. The risk (and cost to customers) of highly proprietary cloud service offerings has already been highlighted by the pain Nirvanix customers felt when this cloud provider went under. Even with some degree of standardization, it’s just plain painful having to move that much data; but without standards, customers are pretty well locked-in. And having gone through many years of negotiating with the large vendors, I have to believe that customers prefer to continue to have the flexibility and safety of a more open environment.
But from the storage vendor perspective, as Yogi Berra said “This is like déjà vu all over again.” As always, “what’s good for me” has been preventing the OpenStack storage standardsfrom developing into a production level offering. For a quick and interesting view on this, read this blog post by Gartner analyst Alessando Perilli.
Meanwhile Amazon has marched steadily ahead, acquiring the dominant market position in cloud storage.
This dominance has been so great that many companies associated with the cloud (e.g., competitive cloud storage companies, gateway providers, hybrid object storage vendors, etc.) have partially or fully implemented a set of interfaces which look remarkably like the Amazon APIs, rather than the “yet to be finalized” OpenStack.
But a new court ruling, determining that parts of Oracle’s Java APIs are able to hold copyright protection, may make it much more difficult for this activity to continue. See Barb Darrow’s recent GigaOM article for a brief review of the reasons why.
So – will this risk be high enough for vendors to step out of their traditional margin-protection comfort zone and work together to create a real alternative? And what will Amazon do? Will they stabilize the use of their APIs with licenses to continue to enjoy the value of the ecosystem that has built up around them? Or do they want to shut this down? Or are they even aware that it matters?
We’ll have to wait to see the next shoe drop.