<img alt="" src="https://secure.bomb5mild.com/193737.png" style="display:none;">

Turbonomic Blog

PSTG FlashBlade: Don't Mind If I Hadoop

Posted by Chris Graham on Mar 18, 2016 9:11:09 AM

pure flashbladeOver the past year and a half, I’ve been fortunate to work in an alliance capacity with a handful of wonderfully disruptive technologies. My first among these was Pure Storage, who at the time offered a single product line, FlashArray, in three flavors: FA-405, FA-420, and FA-450. The FlashArray product line was uniquely positioned as the leader in scale-up, all-flash array (AFA) enterprise storage, differentiated as the most economic combination of performance and efficiency in its category. This past Monday at Pure’s inaugural Pure//Accelerate Conference, Pure Storage announced a product called FlashBlade. And this is where our story begins.

Pure for long has competed with the likes of EMC XtremIO in the mid- and Enterprise markets, and with SolidFire (now owned by NetApp) in the Enterprise market. A long tail of mid-market and SMB players, to remain unnamed, compete with one another, and occasionally against the aforementioned.

Four and five years ago, when Pure Storage (founded in 2009) was gaining steam and EMC had freshly acquired the Herzliya, Israel-based XtremIO, the All-Flash landscape operated on very different tenets than it does today:

  1. All-Flash was assumed to be a niche installation in the data center, and would remain so into the future.
  2. The primary competition for any player in the AFA category was disk – the game was convincing the customer to pay a premium for flash over disk.
  3. Disk economics were just too good to be threatened by an alternative media.

Flash-forward to 2016, and a rapidly-evolved rule set governs AFA:

  1. All-Flash will eventually replace all Tier 1 and (perhaps) 2, 3 workloads
  2. The primary competition for any player in the AFA category – and there are now many – is other AFA players. Every player’s differentiation is its philosophy.
  3. Disk’s destiny is archival data.

Frankly, there are few unique philosophies, and a horde of dime-a-dozen me-toos.

pure storage flashblade rackedWhich is why I admire Pure Storage so much. In the enterprise, there are three and only three players who matter: Pure, (NetApp) SolidFire, and XtremIO. Pure offers the best combination of performance and efficiency, and scales up. SolidFire offers native QoS differentiation, and scales-out. XtremIO offers the best raw performance, and scales-out; it also costs the most.

Pure Storage, publicly PSTG as of October 7, 2015, has had an admittedly tumultuous start. It’s been as high as +23% and as low as -27% relative to its IPO price of $16.01. Still, I’d contend that the swings are at least partially attributable to the startup IPO market at large, and that despite a uphill battle ahead, Pure is positioned to challenge and eventually lead the AFA space for three reasons:

  1. Pure led the philosophical scale-up vs. scale-out charge by boldly stating that for production block storage, scale-up is the way, the truth, and the byte. Scale-up arrays give you the ability to scale performance and capacity independently; meaning your drives are never ‘short-stroked’ – in an AFA setting, think of this as meaning buying additional drives before filling the ones you already own. At scale, scale-out arrays beget waste. Performance, artfully combined with efficiency, is key. As a result, Pure is selling to more and more cloud providers – several of whom I met and spoke with earlier this week at Pure//Accelerate – smartly hedging the cloud adoption bet.
  2. Pure has lower COGS than competitors. Pure reports an ~38% GPM, due largely to its successful incorporation of Consumer Multi-Level Cell (cMLC) NAND (Not + And) flash cells into its arrays. cMLC NAND is the highest-density, least expensive flash breed on the market. Traditionally, it is brittle by nature and excessive read/write operations compromise the chip’s integrity. Pure, however, controls the read/write operations with its bread-and-butter PurityOS – caching as much of the active data as possible, and de-duplicating/compressing the written data an average of 500%.
  3. FlashBlade. Pure is brilliantly diversifying its portfolio. FlashBlade is a scale-out, object-based AFA designed specifically for the Big Data workloads of tomorrow. Scaling-out incrementally at 8TB or 52TB raw, FlashBlade is a first-to-market offering slated to capture share rapidly.
    Back in November, Morgan Stanley projected that Pure Storage would quadruple its share from 1.5% to 6% of the overall $27B storage market from CY15-18, on the premise that it’d continue a 64% CAGR. MS also projected that Pure’s likely introduction of new products ‘in the next few years’ would accelerate this share penetration. Pure delivered in the next few months, and I think the bet they’ve made with FlashBlade will pay-off massively in the long run. NetApp’s Lee Caswell is already dismissing the product a non-competitive. Perhaps today, Lee, but in this case I believe Pure has heeded the Gospel of Gretzki and skated to where the puck is going to be, not where it is today.

Graphic Credit: Pure Storage

Photo Credit: Author

Topics: Storage

Subscribe Here!

Recent Posts

Posts by Tag

See all