The Year Ahead in Data Storage

One of the few things one can foretell safely every year is that there will be an onslaught of predictions articles, blogging, and general prognostication from about mid-December through to about that point in January when we all think it’s got a bit too late to say “Happy New Year”! 

I would hate to buck that expectation!

So, what happens in the data storage world in 2015? Well, having covered the regularity of “predictioneering,” we now come to the next problem – charting a route:

  • One can either go down the “duh, you really had to stop and think about that!?” blindingly-obvious route (a good storage example might be that we will see a continued uptick in the deployment of many types of flash this year!)
  • Or of course, there’s the “if you throw enough mud, something is bound to stick” blind-guesswork effort (“yeah, either HP will buy EMC…or maybe EMC will buy HP”)
  • And let’s not forget the “let’s just make it a great read, because I doubt anyone will check and I will claim credible deniability” plain-and-simple blindness (“scientists based on the Space Station are conducting experiments that will enable exabytes of data to be stored in thimbles-full of condensed thermosphere…that costs almost nothing and could start shipping in October”).

These approaches are all either reassuringly cathartic and/or simply good fun, but I’m not sure that either their inherently overly-specific or overly-grand pronouncements are really very useful. Frankly, we all know that areas such as flash, software-defined storage, integrated/converged platforms, the cloud and object storage are all more or less of interest to vendors and users alike. [If you want to know which are “more” and which are “less” and why, be sure to check out a new ESG Research Report on “Next Generation Storage Architectures,” due to be published in the next few weeks.]   

So, I am going to go down a path-less-travelled and tell you what writing I see on the wall….based on my conversations with literally hundreds of vendors (from big to stealth, and everything in between), channel organizations and end-users. I should warn you – although obviously not apologize – that it’s not quite so dramatic as a checklist of quarterly expectations. This is because the storage ecosystem moves as an amorphous, disjointed mass rather than by sudden orchestrated twists and turns….but nonetheless key trends can be seen from a 64,000 foot level.

The granular-but-important stuff will indeed be the continued increasing use of flash (yes, I said it!), as well as continuing flash-vendor consolidation. Specific technology and architectural changes will also make more, relatively gradual, progress – object storage and integrated-compute (or [hyper]-convergence if you prefer) being obvious examples, but by the way, storage virtualization isn’t a completed job yet either. More users will adopt – and adapt to – the cloud to some extent and in some fashion, and a whole host of new vendors will emerge from stealth; most will sport some association with the term “software-defined storage.” Oh, and the majority of storage vendors will at least learn to spell the word “security” in 2015, because they will find their ability to sell product restricted as the year proceeds if they do not make some allusion to their products being “security-enabled” or the like. [Sneak Peak from this year’s imminent IT Spending Intentions survey – “Security/Risk Management” is the number one business priority driving IT spending – we have never seen this before, and it’s eclipsed things like cost containment and process improvement!] 

But, now, if you mash many of these individual aspects together, you can see that they are all individual streams in a much bigger tidal-wave change: we are heading to an era where storage per se becomes way less architecturally important…..while simultaneously it becomes way more foundationally crucial. Hopefully this doesn’t come across as a riddle! Let me use an analogy: storage has for decades been a “black art” conducted by people in white coats, much as tweaking a car engine’s performance was a matter of cranks, chokes, gut-feel, and trial and error. These days we rely on sophisticated engine management systems that guarantee acceleration when you need it, economy when you want it, low maintenance and incredible ease of use, and that your car will start first time be it in the Amazon or Arctic. For most car drivers most of the time we have moved from needing expert tinkerers to simply being avid users; and so it is going to be with storage. We want – be it the car or storage – a trouble-free, flexible, predictable and economic outcome. Despite dynamic and variable demands we want to say “it just works.”

How we achieve that can vary: we apply a better engine-management system (software-defined storage), we buy a car rather than individual components (convergence) or we rent/lease/Zip-car/Uber to some degree to ensure service-levels (these are the cloud, CoLo, XaaS types of model). Again, whether we are talking a car or a storage system, what matters increasingly is the experience, the outcome, and the results – the what, not the how. But – and this is just as important a realization – is that as we become more removed from the mechanics of the how, so our reliance on that side of things actually increases. This is what I mean by storage becoming more foundationally crucial at the same time as it becomes increasingly subsumed into a hybrid-converged-software-defined-heterogenous-single-global-namespaced-on-prem-cloud solution.

I may be poking some semantic fun at the change but in reality it is huge and the potential benefits are compelling. It can be seen as part of the continued re-casting (or rediscovery?) of IT as information technology rather than infrastructure technology. This change is coming at us in many separate but connected individual elements. Clearly we want apps and results, whether as organizations or individuals. And at the infrastructure level (which of course we still must have…that Space Station suggestion was – shockingly – just made up!) a similar change is afoot. We no longer build home computers and add floppy drives or install SSDs, but we buy preconfigured, almost disposable, laptops and tablets, and use flexible and scalable Google Apps and OneDrive. The data center (or more accurately the business function of IT) will track a similar path.

My prediction is therefore that 2015 is the year when we truly begin to realize, appreciate, and openly discuss that storage is simultaneously less important and more important than it has ever been.  By “less important” I mean that it will not be as prominent in the thoughts and actions of as many IT folks because “it just works.” By “more important” I mean that any time we come to rely more heavily on something (and/or it essentially becomes a utility), it had better darned well work because the impact of a failure is typically more widespread and bad than before, and there are fewer people with - and less institutional knowledge of -  the detailed ability to fix things. In simple terms, making storage less complex to use will demand increasingly complex engineering….of both the software and the hardware: our increasingly data-centric, consolidated, and virtualized world makes anything less than rock-solid as frightening as it is unacceptable.             

Topics: Storage