Datacenter Trends

IDC Compares (Some) All-Flash Array Vendors

More interesting than the results could be the process of trying to evaluate a market that is in constant change.

The industry watchers at IDC have released a new report, "IDC MarketScape: Worldwide All-Flash Array 2015–2016 Vendor Assessment," in which they examine the maturing all-flash array (AFA) market and compare leading vendors in that space. The analyst firm just published a free excerpt from that report, which was authored by Eric Burgener, Ashish Nadkarni, and Eric Sheppard, and I think it's worth reading.

What you think about this report might depend on your reaction to IDC's definition of a "true AFA," which includes only systems that "...have hardware that is unique to the all-flash configuration within the vendor's product line (other than just the choice of populating the array with all flash media) and the array's inability to support hard disk drives (HDDs) must be more than just a marketing limitation ...." The analysts added that all-flash configurations of Hybrid Flash Arrays (HFA/As) do not qualify and were excluded from the report.

In other words, an AFA is a system that uses flash media -- and only flash media -- to meet performance and capacity requirements.

Also, to be included in this study, vendors had to meet a specific set of criteria. They had to have generated revenue from an AFA product line prior to July 1, 2015; to use a multi-controller architecture or a scale-out design with nondisruptive failover; and to use either Custom Flash Modules (CFMs) or Solid State Disks (SSDs) to meet both performance and capacity requirements within the array. They also had to provided IDC with at least two live and available customer references.

The point has been made that these are narrow criteria. The Register's Chris Mellor has called them exclusionary. And to be fair, a number of vendors many of us would consider to be AFA providers did not satisfy the analysts' definitions. It might also be true that a Gartner AFA Magic Quadrant, published last year, was, as Network Computing's Howard Marks suggested, based on arbitrarily defined market segments.

These are, in fact, the at-this-point-in-time conclusions of analysts trying to get a bead on a still-moving target, and they should be taken that way. But I don't see anything arbitrary or exclusionary here. Consider the analysts' explanation of why they excluded vendors that market their HFA/A products as AFAs: "IDC's review of unpublished performance testing data using more real-world workloads (not 'hero' tests) combined with simultaneous data services indicates that there is a noticeable performance difference, particularly with respect to consistent latencies, as a system is scaled up to its maximum throughput range between AFAs and HFAs."

But they are also aware that "this area is evolving," and noted that several vendors are beginning to narrow the performance gap between IDC-defined "true AFAs" and HFA/A offerings. "IDC will continue to keep a close eye on developments in this area," they wrote.

So, who made the cut this time around and why? Here's IDC's list:

The Leaders:
EMC's XtremeIO
HPE's 3PAR StoreServ 7450c
Pure Storage's FlashArray//m )

The Major Players:
IBM's FlashSystem 900/V9000
SolidFire's SF Series
And Kaminario's K2

The Contender:
Violin Memory's Flash Storage Platform

In my opinion, the key insight in this free piece of this report is the analysts' observation that customers are thinking more and more about the potential of these platforms to serve as general-purpose primary storage platforms for hosting a variety of mixed workloads.

"All participating vendors offer solutions that easily support hundreds of thousands of IOPS, consistently deliver sub-millisecond latencies, and can support at least half a petabyte (PB) of effective storage capacity," they wrote. "All of the AFAs from the vendors in this study can confidently be deployed today as general-purpose primary storage arrays to deliver flash performance and enterprise-class availability/reliability, providing opportunities for customers to transform both their IT infrastructure and their business opportunities."

About the Author

John K. Waters is the editor in chief of a number of sites, with a focus on high-end development, AI and future tech. He's been writing about cutting-edge technologies and culture of Silicon Valley for more than two decades, and he's written more than a dozen books. He also co-scripted the documentary film Silicon Valley: A 100 Year Renaissance, which aired on PBS.  He can be reached at [email protected].


comments powered by Disqus

Subscribe on YouTube