EMC’s Xtreme embarrassment

by Robin Harris on Wednesday, 5 December, 2012

EMC’s Pravda, Chuck Hollis, chats in an interview about XtremIO’s problems. This quote from a Storage Newsletter interview may have some transcription errors, since SN is a French publication, but let’s not split hairs. Read it all to learn more.

When will you have a dedicated all-flash array?
We announced our XtremIO acquisition last year. We demonstrated it at VMworld and EMC World. It’s an all-flash design. Made of de-dupe, scale-out architecture, all flash. It fits very precise parts of the market.

Do you sell it?
No, right now we’re in what’s called ‘early access’. Test with the customers, make sure it works right. You’ll probably see it in the first half of 2013 as a general product. It’s got a very specific profile. Random IO/s, extreme performance and tiering do not work. We have a data set where everything has to run fast.

Signal intelligence
Mr. Hollis is paid well to present the EMC party line, and he’s good at it. There are reasons for everything he says.

  1. Pre-announcing the new flash array is intended to freeze the market, especially Violin Memory, Kaminario, Nimbus Data and Pure Storage among others.
  2. “Precise parts of the market” means that EMC doesn’t want customers to see it as a Symm replacement – which given their likely Xtreme pricing and need to keep VMAX billions rolling in – is simple self-preservation.
  3. It won’t ship until the 2nd half of 2013 at the earliest and may slip further. But an H1/2013 announcement is likely.

Unpacking the ship date
EMC is very good at keeping secrets when they want to. In this case they don’t want to.

Why?

By telling us that they’re intending to enter the market for pure flash arrays they are hoping to slow the competitors kicking VMAX to the curb. EMC sales hates losing deals, but since they don’t have a competitive flash array they are stymied.

If the product were shippable in Q1/2013, EMC would deploy evangelists, including Mr. Hollis, to give NDA presentations, offer test lab visits, early delivery units and so on. But the “early access” program – “test with the customers, make sure it works right” – is a beta test. Normally that would come late in the development process when announce and ship dates are reasonably firm.

The dates aren’t firm if “probably” H1/2013 is the best EMC can offer. When XtremIO was bought, EMC predicted a Q1/2013 product launch.

The StorageMojo take
XtremIO hadn’t shipped a beta when EMC bought them. It’s clear that XtremIO’s “great technology” is further from a “great product” than they’d hoped.

Thus the damage control effort.

To limit embarrassment EMC may announce at EMC world in May with “planned availability” for Q3. But given the dynamics – flash arrays have a lot of interesting wrinkles – Q4 is my forecast for shipments of a de-featured product.

What went wrong? Under-estimating market demand for all-flash arrays plus over-estimating VMAX with flash cache appeal, resulting in a reluctance to pay $2B+ for a less-mature-than-they’d-like, but installed product company like Kaminario or Violin.

The year’s delay will be costly for EMC as flash array competitors become firmly entrenched in major accounts. Fixing that will take more than spin.

Courteous comments welcome, of course. I’ve done work for Kaminario and Violin, and I’m still a little peeved over EMC threatening StorageMojo.

{ 13 comments… read them below or add one }

Chuck Hollis December 5, 2012 at 5:26 am

Robin

I was very dismayed yesterday to see that Storage Newsletter made a hash of a recent interview. You can tell by reading it that they had multiple and severe translation challenges. And, of course, they never bothered to send along a draft back to me for verification — which is standard practice in these cases.

We’ll be talking with them about that.

I was doubly dismayed to see that you had grabbed a snippet of their poor work, and decided to make it all your own.

To be clear: the Xtremio product is doing quite well, thank you, and meeting all of our expectations. We’re looking forward to a public GA in 2013.

And, at that time, we’d be glad to share more information with you.

— Chuck

Calle December 5, 2012 at 12:52 pm

Just wanted to point to the fact that it wasn’t EMC that said Q1, according to your linked article, I quote:

“Chris Evans, consultant at Brookend, told Computer Weekly: “I think EMC’s purchase of XtremIO was probably for talent acquisition rather than technology.”

“EMC have already stated that there will not be a beta of Project X until the end of this year, with GA product in Q1 2013, so clearly there’s no hardware platform to start shipping.””

Robin Harris December 6, 2012 at 2:12 am

We’ll know for sure on July 1, 2013.

Mike Chudzik December 6, 2012 at 9:55 am

I don’t think there’s much left to debate on the timing other than when in 2H2013 we’ll see general availability of XtremIO. Chuck himself confirmed this on his blog yesterday.

Chut December 7, 2012 at 10:09 am

Do you really think Violin and Kaminario are worth $2bn?? I think they should come down to a more reasonable number if their endgame is to get acquired.

I would agree with their quote if they are planning to stay in the game longterm.

nate December 7, 2012 at 3:12 pm

I’ll be curious to see how far out this Xtreme IO stuff can scale once it comes out. I read one claim that said it can only go to 8 nodes or something.

I remember when EMC announced their VMAX and had grand plans to scale that out as well (to something like dozens of controllers and millions of IOPS). But here we are 3 years later and they haven’t gone beyond the original 8 controllers.

I asked 3PAR a few years ago if they ever plan to go beyond 8 nodes – their architecture supports it – it’s technically possible, probably not even all the complicated. But they said no, no plans to go beyond 8. The main reason was risk, they designed their clustering system so it is not tightly integrated(unlike VMAX) so a failure in one controller is less likely to impact the rest of the controllers in the system (clusters are bad like that).

With their last high end refresh 3PAR chose to double the power of the controllers rather than double the number (EMC refreshed their VMAX not too long ago I think and increased the power of their controllers as well). But even with that designed in that safer manor, I’m sure it doesn’t eliminate the chance, and the more systems participating in the cluster increases the likelihood of failure.

NetApp made a bunch of noise about their scale out cluster stuff earlier in the year but once you dug into the details it wasn’t nearly as nice as it was pitched. It was more like a workgroup of arrays rather than a real cluster.

Scale out traditionally seems to have a trade off of throughput vs latency from what I’ve seen in the NAS space at least.

So what sorts of compromises will EMC have to make for this scale out XtremeIO stuff..hmm

Jean-Jacques Maleval December 9, 2012 at 8:45 am

Robin,

You write: “This quote from a Storage Newsletter interview may have some transcription errors, since SN is a French publication, but let’s not split hairs.”

Which transcription errors did you find in the interview. I’ll be happy to send you the full audio recording. Chuck just didn’t like the transcript and would prefer a “translation” of his real words.

Cordially,
Jean-Jacques

Corentin Béchade December 9, 2012 at 9:19 am

I’m sorry Mr. Hollis you can’t blame us on that one, the part of the transcript M. Harris choose is exactly what you said in the interview, you even wrote it down in your own version of your own interview.

And the analysis is about what you said, not about any possible mistranscription. Stop accusing us and our “poor work” every time an article point to our interview, we cannot be held responsible for things you said.

Plus Mr. Harris has the honesty of pointing out that there may be some transcription error, even if it’s not the case here. We did not add anything that you did not say. It doesn’t make sense to accuse us, Storage Newsletter, for what you said. This is not an error we did.

PS: My apologies to Mr. Harris for getting fired up about something that wasn’t really the point of your article.

– Corentin Béchade

Bacon Weezy December 10, 2012 at 8:35 pm

I completely agree with Robin! However, if you look at how EMC responded to Greenplum bombing (spinning it off with other under-performing assets from VMware – http://bit.ly/SL2ePr – instead of writing off the losses like HP did with Autonomy) it’s obvious they can only handle one bomb at a time.

David Nicholson December 12, 2012 at 10:34 am

I would argue that the most educated shopper walking up an down the aisles at the “All Flash Array Super Store” was EMC. With pockets full of cash, EMC had the opportunity to evaluate every single flash array start up on the face of the Earth. EMC put it’s money on XtremIO.

Although I am sure Violin scored highly in the “How bright is your sign and can it be seen from the 101?” category…and companies like Pure took the “Funniest video with a rocket and a dinosaur.” award…EMC seems to have made an investment based upon different criteria.

(Did you know that the monkey in that Pure video was more expensive that the hilarious actor?)

I don’t think it is irrational to assess the all flash array TAM and make a bet on tech that will roll out at a later date. Certainly not more irrational than $2B for Kaminario or Viloin. Even shrink wrapped into a two-for-one pack.

We sure are a long way from Margot, aren’t we Robin? ;-)

Happy Holidays,

Dave
lesser known shill for EMC

Patric January 14, 2013 at 2:31 pm

It is rather amusing to read through a thread full of accusations and blaming . What is all this about?. A company aquires a start up and a product or to be exact IPR , and existing technology which will be streamlined to fit the corporate technology standard . It builts up an execution machine to deliever a product and service which meets customers and shareholders and employees expectation. All the rest is very poor speculation and let us just summarize by the end of 2013 which company rocked the flash market the most ? As far as I remember EMC never aimed for 2 nd. A good start into 2013 Patric

John Martin May 4, 2013 at 5:44 am

XtremeIO may have been a genuine bargain for EMC. Their tech would have been at a point that was reasonably ready to commercialize which, even if its a complete bust, the “mistakes” and lessons learned would probably accelerate EMC R&D efforts in what might be a very disruptive market, by at least 12 months.

Add that to the likelihood that the money spent would have come from profits earned outside the US, held in low tax countries (e.g. Ireland), invested in another country where the profits from R&D are taxed at an even lower rate (Israel) on a bunch of really smart guys who participate in one of the most successful start-up regions on earth outside of Silicon Valley that has a currency that fluctuates very little vs the USD.

That is the kind of R&D that makes the bean-counters happy, both those inside EMC, as well as those members of the financial community who can have significant impacts on stock price movements.

With that taken into account the XtremIO purchase makes sense whether it pans out or not … and you never know “maybe the horse will learn to sing”

Thomas Leavitt November 26, 2013 at 3:59 pm

… and the winner is?

I see a press release from EMC announcing GA for this product as of November 14th, 2013.

Thomas

Leave a Comment

{ 1 trackback }

Previous post:

Next post: