Solid State Storage Ready For Prime Time

There's a well-worn saying in the storage world to describe new and emerging technologies, when no one is quite sure what the actual adoption rate will be: "There's more written about it than on it."

In many ways solid state storage (notice I didn't only say SSD, and that's deliberate) has felt a little like that over the last few years. There's been lots of talk and some amusing puns ('spun up about solid state', 'solid results', 'no-spin zones' etc.), but is this solid state stuff really destined to matter in the big scheme of things? More narrowly, we all know solid state matters in consumer electronics, but will its abilities--performance, compactness, low power usage--translate into the enterprise market? I'm here to state (no pun intended!) that the answer is a definite yes, although the motivations behind the adoption may be broader and different to what's expected. Solid state storage is headed for, and arguably already in, prime time.

So what does "prime time" mean? Well, for starters this is about market relevance, not a take-over per se--spinning disk has plenty of life in it yet. But here's the point--even terms like "relevance" and "take-over" are heavy with historical bias, for we've become accustomed over decades of only talking about storage from one perspective: capacity ("how many $ per GB?"). We need to start thinking differently--all users have stored data that's active, demanding relatively more I/O, and also stored data that's inactive, demanding relatively more space. And, whether it's I/Os or TBs of capacity that are needed, the one consistent factor for IT is that all storage needs to be as inexpensive as possible. Put differently, solid state storage will only be relevant if and where it makes economic sense.

Performance v. $$

But surely solid state is all about performance? Yes and no. We've always known there is a performance challenge resulting from the physical inability of traditional spinning disks to serve up more than a few hundred IOPS each, yet we've found myriad ways around that challenge even without the benefit of solid state storage. So, in fact, even the performance of solid state is not, per se, the root value of solid state as many seem to think. It is not that we could not achieve performance before, but rather that achieving it was way too complex and expensive. In other words, the performance of solid state is--at least for the vast majority of users--a means and not an end in itself. Where solid state is becoming relevant also has nothing to do with capacity (again, at least for the vast majority of users) and has everything to do with serving I/O economically.

With just about every storage systems vendor supporting, developing, and promoting solid state storage in one way or another, it's no surprise that market adoption is happening--but the move from solid state being a specialist, marginal tool for extreme applications to it being a generalist, prime-time technology is based on the growing understanding that--counter-intuitively for something that seems so expensive--it can help to control costs. Of course, the fact it does that without negatively impacting performance, and even often improving it, is great. But its adoption would soon stall were it not economically attractive. Sometimes this fact is hard to discern through the plethora of hyperbole and "specsmanship" that many solid state providers are promoting. You would think that everyone is capable of using a gazillion IOPS! But it's a bit like advertising the top speed of cars--it conveys a degree of capability even if very few drivers ever truly use it or need it.ESG Research

So, how far are we into this prime time phase for solid state? Is this column itself just hype or is solid state really becoming relevant? ESG recently conducted primary research and discovered that just over a half of the enterprise-class end-users surveyed either already have solid state in some form (34%) or plan to have it by mid-2012 (17%).

As expected, the percent of adoption far exceeds the share of the total storage market--whether measured in revenue or total capacity--that is represented by solid-state storage, which is neither surprising nor likely to change anytime soon. At least for the foreseeable future and for most users, solid-state storage is something that is used in small amounts for big effects. Digging a little deeper into the adoption statistics, the research found:

-- Solid-state drives (SSDs) are the most common solid-state storage implementation type. Looking ahead, standalone solid-state storage appliances--whether serving as primary storage or a cache--appear poised to make the most significant gains over the next year and a half in terms of market adoption among both current users and potential adopters.

-- While performance remains the most mentioned adoption driver, reliability is also now being recognized as a key consideration for solid-state adoption. Indeed, while 36% of respondents identified performance as the primary reason their organization deployed solid-state, nearly two-thirds selected something other than performance. On the flip side, raw cost and some perceived technical immaturity are the top perceived inhibitors for non-adopters.

-- In terms of satisfaction, the vast majority (ranging from 78% to 92%) of users are either 'satisfied' or 'very satisfied' with their solid state purchase across a range of key criteria.

Adoption outlook

Given that just about every storage vendor--and a whole host of emerging start-ups--is actively promoting the benefits of solid-state, it looks like a foregone conclusion that the adoption of the various technologies under the "solid state storage" umbrella description will continue to ramp up over the next few years. The options can be segmented into three main infrastructural implementations: in the server, in the storage subsystem, or as a standalone appliance/array. There are also two main usage options: persistent storage (a tier) or temporarily stored copies of data (a cache). And, of course, there are software [only] tools to manage and optimize the use of solid state. As users gain more knowledge of solid-state's capabilities, and as the functional and financial value of the vendor offerings increases, the IT community itself expects adoption to increase: nearly two-thirds of respondents in ESG's research indicated that they currently view solid state as a niche technology, but expect it to have widespread applicability in the future, while a further 15% believe it has already achieved that status.

The adoption of solid state will happen because it makes economic sense; it will grow as a prime-time I/O platform, but is unlikely to make much of a dent in the overall traditional spinning disk market in terms of volume/capacity, where it does not yet make such economic sense. There will however be a growing market beachhead for all sorts of solid-state implementations--all the way from "turbo-boosts" on existing systems to stand-alone flash-based arrays (whether 100% flash or hybrid). The bottom line is that, on average, a relatively small amount of solid-state capacity will soon be the norm in almost all user environments ... where it will be serving a relatively large percentage of I/O, and where its performance will be measured at least as much in terms of $$ as in terms of speed.

Mark Peters is a Senior Analyst at the Enterprise Strategy Group, a leading independent authority on enterprise storage, analytics, and a range of other business technology interests.

It's time to get going on data center automation. The cloud requires automation, and it'll free resources for other priorities. Download InformationWeek's Data Center Automation special supplement now. (Free registration required.)