« Why Oracle Doesn't Like VMware | Main | Oracle Wins Virtual Iron, Customers Lose »

May 08, 2009

Comments

Alex McDonald

"We'll wait until we see it to offer specific comments, but I don't believe there's any way NetApp could ever afford this level of investment."

The question is, can EMC? Or more to the point, EMC's customers?

http://blogs.netapp.com/shadeofblue/2009/05/pay-more-tax-and-retire-at-70.html

Chuck Hollis

Hi Alex

Your logic is flawed, if you think about it.

You're implying either (a) customers don't receive value from this investment, or (b) we wasted all that R+D money.

I don't think either is true.

Speaking of investments, I forget how much NetApp paid for Spinnaker ... it's been so many years.

I have a hazy recollection that perhaps it was in the hundreds of millions? How about that Topio stuff that never went anywhere? I forget how much that acquisition was, too.

And no signficant revenue from either.

Now that, Alex, is expensive!

-- Chuck

Hiro

A "Y" is not the symbol for Yen. The symbol for Yen is "¥". In some places, confusing the two is considered culturally insensitive. I guess you knew that.

ヒロ

John F.

Hi Chuck,

I still don't quite get it. First you said NetApp couldn't make such an investment, and now you're saying they did in Spinnaker? NetApp did buy Spinnaker in 2003, and released and has been selling Ontap GX since 2006. If you need an update on the status, see http://www.theregister.co.uk/2009/01/26/netapp/print.html


Now, back to the subject. So you built 750 V-max system, and 80 went out to beta testers. Are the other 670 sitting in a warehouse or what?

Ok, by your own admission, smaller companies CAN make big bets too. Other than contradicting yourself, it still doesent mean much. It's not the size, but the outcome, of the bet that matters. I guess all you're left with is that EMC is a really really big company. I won't argue with you on that one.

John

Chuck Hollis

Sorry, Hiro, I just cut and pasted from the article I saw. Besides, I don't know how to insert the character for Yen.

-- Chuck

Chuck Hollis

Hi JohnF

Let me clarify the comment.

NetApp spent $300m on Spinnaker in 2003. For many years, NetApp promised a "converged OS" which has never materialized, and may never do so -- as pointed out in your link.

You are correct, NetApp finally *did* ship an incompatible clustered NAS product in 2006, but it appears to be somewhat unpopular with customers.

For example, very little is ever said about this product on NetApp earnings calls. Most analysts don't track it any more. We don't see many press releases from your PR machine.

Dual-controller filers we hear a whole lot about, though.

I was hoping to differentiate between a multi-hundred million dollar investment that results in a successful product with significant market share -- and a multi-hundred million dollar investment in a product that apparently has failed to gain any traction over several years.

The same sort of "expensive failed investment" comment can be made regarding NetApp's $150m Topio investment in 2006, which recently suffered the same fate.

Or perhaps the 2005 acquisition of Decru for $275m which also appear to be flatlining as well.

You're welcome to google any of these to find the breathless press releases from NetApp at the time. They're very entertaining reading.

Alex was appearing to criticize EMC's investment in Symmetrix V-Max as unnecessarily expensive.

What he hopes to gain by positioning it as such is beyond me -- it's a very weak position to take in front of customers and the industry.

Great products in our industry usually require serious investment -- there aren't many shortcuts. I think most people understand that.

But I guess competitive bloggers have to work with whatever they've got at hand.

I also wanted to point out to him (and to you) that what is really expensive is massive investments in M&A and subsequent product engineering that result in market flops.

For example, I would point to the $725m of NetApp acquisitions outlined above, and offer that NetApp has very little to show for it.

Does the clarification help?

-- Chuck

Vadim Solovey

Chuck,

I just love to read this ongoing XIV bashing. We have 3x XIV systems for last 2 years, stable as a rock and fast as nothing we had before.

Chuck - please, - a little more balanced approach will make your blog less marketing buzz,


Chuck Hollis

Hi Vadim:

Actually, it's not "bashing", it's reporting the facts.

Most people in the industry (as well as many customers who have evaluated the product) consider XIV a failed effort, and they are saying so publicly. As am I.

We heard that there were a few companies in Israel that got complementary or discounted XIV arrays from the development group in exchange for saying nice things about the product.

Was your IT consulting group one of those?

Or did you do a thorough evaluation of the alternatives on the marketplace, select the best, and pay for it with your own money?

If you are saying nice things in exchange for free hardware or some other consideration, we would call that "pimping". It's not a nice word.

I'm glad your arrays aappear to be running well. That doesn't mean that everyone is in the same situation.

And -- whatever you do -- be very careful when you replace a failed drive!!

-- Chuck

Sajeev Aravindan

Microsoft spent 6 billion dollars to develop Vista, but that doesn't seem to interest customers, does it? Making an argument entirely on the investment made appears to be a flawed to the core argument as well as assuming that others need to make a similar investment to catch up.

May be they don't need to, may be they know a better way to do this like having a single platform which can run standalone as well as scale well. Why spent on coming up with a new product if your existing product can be made to scale up. Wasted efforts can be better utilized.

Chuck Hollis

Hello Sajeev (from NetApp)

I would say that not all big investments make big impacts, but most every big impact requires a big investment.

The second part of your comment is a bit hard to follow, can you clarify? I would point out, however, that what's unique about the V-Max architecture is that it introduces an entirely new scaling paradigm not seen before in our industry, which directly addresses the "make a small array bigger" question in a new and useful way.

-- Chuck

Sajeev Aravindan

You conviniently seem to forget Ontap GX, Isilon or Panasas. There is probably something new in V-MAX but scale out is not certainly new. May be you can point me to a place that talks about this new scaling paradigm.

Geoff Mitchell

A great investment indeed, Chuck. Considering the challenge to sell even the current DMX models in this climate, some would consider this akin to trying to sell a Duesenberg in the middle of the Great Depression. That takes brass ones and good for EMC for stepping up on behalf of the industry to push the high end, when there are customers out there with petabyte issues that cannot find storage today with sufficient scalability or throughput.
Symmetrix storage will continue to be the most expensive on the market and with good reason - if you're betting the business on your buying decision and you need the capabilities of this platform, there is nothing that really compares.

Now, referring to the voice in my other ear...
That EMC makes significantly more investment in R&D than its competitors is interesting but not necessarily significant. If this was such a major factor, Equallogic, 3PAR, Compellent and several other traditional tier-2 vendors would not have increased market share over the last few years (according to IDC) and the Symmetrix market would be flourishing.
Secondly, with the billions that EMC has invested across its product line for the last 8 years would have sent its major competitors packing. It's not the case.

Will EMC sell on volume, or will EMC sell on price? Future 10Q's will indicate this, but I have to think that it will be quite a while before EMC's R&D investment is repaid on this model. Reducing EMC's cost of sales is going to be key.

[disclaimer: I don't work for EMC, a competitor, EMC partner or reseller]

Storage Pepper

What I find sad is that V-Max doesn't do anything for EMC in the HPC space. For all the pooplah about "real" fc, it's not what the high end cares about now. When a customer has a 5k, 10k, or 20k+ cpu core HPC cluster they want NFS. For high end NFS EMC has nothing to offer.

V-Max + "Cloud" storage is a marketing smoke screen, meant purposely to distract customers from that fact and other weaknesses (dedupe comes to mind). Chuck rips on NetApp's GX, but these days I never see EMC at the HPC table. Customers ask me about Isilon, GX, Panasas, sometimes GFS and others. Almost never EMC.

And no, I'm not a NetApp employee Chuck. Although I've read they're a good company to work for...

Alex McDonald

Chuck, my comment didn't seem to get through last time due to some TypePad problem, so here it is again.

Topio wasn't so great. Yes, we gained some clients, some revenue and some technology, but going forward, I think NetApp took the right decision that it didn't make financial sense to keep investing in it. Some you win, some you lose.

As to Spinnaker, that was an investment in technology that is paying dividends already (ONTAP GX) and will play a much larger part of ONTAP going forward. That's one where we invested well.

The amount you spent on V-Max -- what is effectively a commodity hardware system -- is, frankly, obscene. The reason you felt compelled to spend these mountains of cash is the fragmented nature of EMC's storage systems.

Centera. How much development is this platform going to get now that the Belgian development site is closed?

CLARiiON. Some hardware changes, but the software stack is stuck in the "Real LUNs" mindset. How much investment did this one get?

Celerra. Some changes in the last few years, but the platform especially the software) is in desperate need of an upgrade. Bolting bits on to the FrankenNas doesn't seem to me like a route forward; the suggestion that it's the foundation of a "unified storage system" is laughable. How much investment did this one get?

Atmos. Yet another technology, this time focussed on cloud. How much investment did this one get?

V-Max. Huge amounts of investment, that we do know. And for a system that's based on commodity hardware, the software must have taken the lion's share.

Yes, some of the V-Max RD will leak over into the other products eventually. But what suffered in the other product lines while all this was going on?

If the answer is "nothing suffered", then the total development, qualification, testing bill for EMC must be very large indeed.

Or, as I suspect is the case, each product line got less than what NetApp spend on Data ONTAP. Effectively, EMC are operating as a big umbrella company with 5 or more small and independent storage subsidiaries. Some of them would appear to be smaller than 3Par.

PS. When is FAST out? Or are you indulging in a little bad marketing? http://chucksblog.typepad.com/chucks_blog/2006/12/netapp_bad_mark.html

Chuck Hollis

You're right -- V-Max is probably overkill from an architectural perspective for HPC environments.

These HPC environments usually don't need big write caches, bulletproof availability, support random I/O profiles, advanced replication, etc. that enterprises desire.

Usually much more modest storage kit combined with a clustered file system is more than enough.

Frankly speaking, we never saw much of a market there to develop a differentiated product specifically for that space. Sure, there are a few big environments, but -- in aggregate -- the HPC segment didn't seem interesting enough to warrant any focused R+D.

We tend to use CLARiiON with file systems like IBRIX and a few others when we have a customer that wants that sort of thing.

So, it's not surprising you rarely find us at the HPC table -- we never really intended to go there.

Thanks!

Dave

Good post and comments - glad my RSS feed is back up I was missing all the fun.

What people seem to be ignoring in my view is that VMax could be the 'last man standing' in the high end. Just like when IBM out-invested the competition making it more attractive to run 'the next app' on mainframes. Hitachi bailed, Fujitsu bailed and IBM now has the market to itself. IBM invested in CMOS and Sysplex and a bunch of software that Scott McNealy called a 'hairball.' It's quite a profitable hairball today.

VMax's CMOS analog is Intel chips and like Sysplex it offers the ability to do things at greater distances. FAST (when it ships) should make it more attractive to put 'the next TB' on VMax.

Is it worth some premium for Symmetrix customers to store data on VMax and get the highest reliability, the best remote recovery, etc? I would think so, especially if they don't have to rip out all their Symmetrix processes.

Hope to find out next week at EMC World.

Chuck Hollis

Hi Alex:

Some quick answers ...

EMC continues to invest in Centera. It currently de-facto defines the CAS (content addressed storage) market, and has no serious competition in the marketplace. Market leadership requires sustained investment.

EMC continues to invest in CLARiiON. It currently is the market share leader in mid-range (dual controller) storage, per IDC. Currently, its feature set, performance and availability attributes are unmatched by other vendors. Market leadership requires sustained investment.

EMC continues to invest in Celerra. It currently is the market share leader in NAS, per IDC. Market leadership requires sustained investment.

EMC continues to invest in Atmos to help define and grow the COS (cloud optimzed storage) market. Exact numbers are still confidential, but we're happy with the initial traction. Market leadership requires sustained investment.

EMC continues to invest in Symmetrix V-Max. The previous product architecture (DMX) was the market share leader, per IDC. V-Max shows every sign of widening that gap. Market leadership requires sustained investment.

So, by comparison, are there any categories where NetApp would like to claim market share leadership, per IDC? Or perhaps claim innovation in entirely new categories of storage (e.g. CAS, COS)?

You are welcome to your opinions, but one thing you can't argue with -- our investment strategy has resulted in broad market share leadership from EMC in almost every storage category, plus innovative work in entirely new storage categories.

Market leadership requires sustained investment.

And when FAST ships, you'll be the first to know!!

-- Chuck

Bobby Asghar

Chuck,

Great post about EMC's muscle to invest in technologies that it wants to. Besides some contribution to the cloud/virtualization thing, the way I look at V-Max is; it's EMC's attempt to let mid-enterprise and above customers get a "Symmetrix" for the same buck they pay for high-end Clariion. In this sense, it may be a good way to beat NetApp in some cases where customers end up choosing FAS6080 over CX. Very good move, in my opinion.

I got interested in some of the cross blogging. This post is probably saying on behalf of V-Max, "My daddy is bigger than yours"...Try if you can match it :-)

By the way, the nature of NetApp's investment in ONTAP (or ONTAP GX) is different than EMC's investment in V-Max for the primary reason that NetApp is a unified platform architecture. I am sure V-Max will continue to serve the higher-end of the market, while NetApp continues to gain territory wherever it can.

Finally, it's kind of hard to predict market's response for the enterprise cloud concept yet. V-Max, with its inside-the-box virtualization, may be the first right step towards cloud model from the back-end. I liked the way you explained about it somewhere on your site though.

Thanks,
Bobby

Chuck Hollis

Hi Bobby

Very insightful comments, so thanks!

And, yes, there is some serious cross-fire between some very competitive companies on these blogs.

All part of the fun, IMHO.

Thanks for writing!

Nigel Poulton

Chuck,

I hope you dont mind me cutting and pasting a comment I left on Alex McDonalds blog at NetApp (I first noticed this little blogwar over at Alex's blog as you usually create content faster than I can read it....) Anyway -

As an HDS specialist (I dont work HDS, Hitachi, HP..... or EMC for that matter) I am impressed with, and a little jealous of the V-Max. I half wonder whether Hitachi will be able to keep pace considering the current economic climate et al.

In the type of work I do, high end storage for mission critical systems, I see Hitachi storage up against EMC all the time, with the occasional mention of IBM. Rarely, if ever, is NetApp brought to the table (Im not trying to dis NetApp).

One of the reasons Hitachi and EMC are the platforms of choice is precisely the point being debated - their kit is seen to be the "best", most feature rich, and most reliable. That reputation doesn't come by cutting corners with R&D, testing.......

If EMC had scrimped and developed the V-Max on a shoe-string, so to speak, then it would never even be considered by the companies that I do work for.

I also keep hearing people say that fortune favours the brave etc during economic downturns, and that those who invest during these times usually come out stronger than those who pull their investment budgets..... Im not old enough to know how true this is but I see the principle behind the notion.

BTW Im not sure it works the same for governments, who, by design, dont/shouldnt really be revenue generating machines.

I for one hope that Hitachi are investing similar amounts in their storage line otherwise I might end up being a specialist in a line of technology that finds itself lagging a long way behind.

Just my penny's worth.

Chuck Hollis

Hi Nigel

I agree with you.

Many people in this industry (NetApp included) really don't fully grasp what the high end is all about, and they tend to disparage or discredit what they don't understand. I think they end up showing their ignorance in many cases, which is not pretty.

You're right -- EMC's game plan is to pour on the R+D during the downturn, and (hopefully!) reap the benefits on the upturn.

Part of me sort of hopes that Hitachi is investing in something big and shiny. Right now, EMC is clearly running away with the high-end game, and while that feels good short term, it's not good for the long term.

That being said, I don't think Hitachi is making that investment.

And that's too bad for our industry.

-- Chuck

SANd Man

These are impressive numbers, Chuck. And while they may speak to the scope of EMC’s resources, I don’t see how they automatically translate into real innovation.

There are many in the industry who believe – rightly so, in my opinion – that the real innovation is being done by the smaller industry players. Industry players who maybe don’t have the budget to bankroll senior executives who spend their days trashing competitors via corporate blogs. EMC has encountered a few of these nimble, innovative young companies through the years, and has recognized the threat they present. EMC promptly acquired them.

OF COURSE EMC has massive resources at their disposal. Symmetrix has been a strong market performer through the years, and aggressive marketing and sales tactics have enabled EMC to corner the high end market. But I look forward to the day when one of these nimble, innovative upstarts rises up to humble EMC - and resists the urge to sell-out. Customers, competitors, and partners alike - we’ve grown tired of EMC’s hubris and bluster. We’re itching for a new, credible alternative to Symmetrix. It’s just a matter of time until we get it.

Nevermind a new discussion about all of the “next-generation” innovation that’s baked into V-MAX. EMC has opinions here, of course. Competitors and customers have their own opinions…

shiningarts

I am not in a position to interject my two cents to the storage space, since my expertise lies in content management area. However, I do have a single cent to contribute here since I have been watching the blog and considering the pros and cons as everybody has been weighing in. It is in our human nature to boast or brag about what we do well, when we think we are the best at something, or simply to complain that we could do something better than someone else. However, nature and history suggests that we need to display a modest temperament and humility in order to better understand the situation that we are facing and to rationally assess what is going on around us. One can complain and diatribe about blogging or gobbling up small companies. In reality, these kinds of actions are not particularly unique to the technology sector; contrarily, they have become trendy fashion or even considered as the greatest achievement in these days. The political juggernauts are employing these tactics and gobbling up private companies, big and small, with just the promise of payment in future generations. When EMC does it, at least they are using their real monetary collateral. Interestingly, it seems that a majority of people approve of these tactics in these days. In fact, incongruously, those people who are complaining about this sort of tact in this blog might have already approved of it in another forum. The decision to choose what kind of storage device you are going to employ is depends on how much damage you are willing to tolerate. Mission critical projects cannot afford to use moderately priced storage devices because they cannot compromise quality over price. Every product has its proper usage and tolerance level for its intrinsic purposes. It is not very helpful to hurl insults between each other’s product since innocent bystanders can get hurt in the crossfire. Rather, it would be advisable to discuss each other’s merits positively in order to advance everybody’s welfare and learn from it.

Nigel Poulton

Chuck you should probably tell SANd Man about the USP V.....

After all, he did say "...We’re itching for a new, credible alternative to Symmetrix"

:-D

nate

Since this is the most recent V-MAX post I suppose this is a good a place of any to ask this question.. I was reading another blog this morning(forgot which, it's on another computer), I think another EMC guy who was going through the various specs and stuff of the V-MAX and the V-MAX SE(?)..

The V-MAX SE has 1 storage engine vs the V-MAX big daddy has (up to) 8. Is there a seamless upgrade path for SE users to go to the V-MAX? I'm just thinking since you have a more "independent" interconnect(not a passive back plane like 3PAR even though the back plane is twice as fast), in theory at least it should be pretty easy to go from a V-MAX SE to a V-MAX(e.g. no downtime, no data migration etc). Is the storage engine behind the V-MAX SE the same as the V-MAX(as in identical part for part)? Is the only difference the additional virtual matrix stuff? The V-MAX has it the V-MAX SE does not?

Have your guys come up with some more in-depth technical architecture docs yet? I haven't gone looking in depth recently but I figure you'd be on top of it if/when they get them.

I thought the recent vmware benchmark from a couple days ago illustrates how storage arrays of the future have to be re-thought about probably from the ground up to try to re-architect for the 100x improvement in IOPS that EFD gives. It's kind of crazy to have an entire storage array for just 10 EFDs. It's not a knock on the CX4, as I believe all traditional storage arrays have the same issue, V-MAX, even my own 3PAR included(which I extrapolated from the vmware numbers could drive 13 EFDs with 4 controllers, that's 8U of controllers for 1U of disks). EFD seems to have snuck up on the industry, a bunch of vendors are touting it, but the silicon required to drive them effectively just isn't there yet.

It'll be interesting to see what the industry comes up with in the next 5 years or so to deal with EFDs in a more efficient manor.

Joerg Hallbauer

OK, I know I've bben around a long time, and sometimes my references are old, but, has anyone (Chuck?) read a book called "The Mythical Man Month"? Long story short, it deails the development at IBM of a new OS way back in the old days. One of the key take aways from the book is that adding more programmers to an already late project just makes the situation worse, not better. Bottom line is that telling us that something as complex as the V-MAX having it's development spread across a lot of different countries, and using thouands of people doesn't increase my confidence that it's going to be bug free. Just the opisite is true, I suspect that the bigger the team, the more bugs will be introduced. Overall, I think that some of EMC's customers are going to love their new V-MAX, but there will probably by a few who will rue the day they decided to buy a V-MAX.

--joerg

The comments to this entry are closed.

Chuck Hollis


  • Chuck Hollis
    SVP, Oracle Converged Infrastructure Systems
    @chuckhollis

    Chuck now works for Oracle, and is now deeply embroiled in IT infrastructure.

    Previously, he was with VMware for 2 years, and EMC for 18 years before that, most of them great.

    He enjoys speaking to customer and industry audiences about a variety of technology topics, and -- of course -- enjoys blogging.

    Chuck lives in Vero Beach, FL with his wife and four dogs when he's not traveling. In his spare time, Chuck is working on his second career as an aging rock musician.

    Warning: do not ever buy him a drink when there is a piano nearby.

    Note: these are my personal views, and aren't reviewed or approved by my employer.
Enter your Email:
Preview | Powered by FeedBlitz

General Housekeeping

  • Frequency of Updates
    I try and write something new 1-2 times per week; less if I'm travelling, more if I'm in the office. Hopefully you'll find the frequency about right!
  • Comments and Feedback
    All courteous comments welcome. TypePad occasionally puts comments into the spam folder, but I'll fish them out. Thanks!