'Meh...it's only a Billion Dollars...'

NetApp worry me as a company; despite their record revenues this quarter, they strike me as a company in trouble. And as an end-user who wants/needs a competive storage market, this is a little concerning.

Now obviously, you are now thinking that 'Bod has gone mad, so I better explain my reasoning.

Over the past year or so, NetApp have been quietly dropping products under the guise of focussing more on their core but if you look at things, their core is actually very narrow. Dropping unprofitable lines is obviously generally good business but I don't see these lines being replaced with anything. Their product range is narrowing, this is not the actions of company confident in their ability to provide solutions to a market-place which will become increasing solution orientated. This is a company who is willing to be a little cog in the grander scheme of things!

The struggle to get OnTap 8 out of the door has in my opinion meant that the company has not really focussed on providing innovative new products. NetApp are currently not innovating and the rest of the market are catching up and some could rocket past them.

Some of the comments I've heard from people who have looked in more detail at OnTap 8 are concerning as well. If you are running OnTap 7; it appears that there is little for you, if you merely upgrade. It sounds like that a full re-implementation is required to take advantage of features like 64 bit aggregates.

When EMC announced Atmos, NetApp dropped big hints that they had a RESTful object oriented storage product in the works. This has yet to surface and I've not heard anything more than 'watch this space' muttering but there's no product shipping.

Another reason that I am concerned about NetApp are that they are a single product company; if OnTap 8 struggles to gain acceptance, there is little for the company to fall back on. And there seems to be little appetite at the moment to broaden the NetApp product range and as I said earlier, they are indeed shrinking their portfolio.

The failure to take over Data Domain and loosing that battle to EMC, I suspect damaged the company's confidence internally and I wonder whether they currently have the appetite to embark on an aquisition campaign but surely that is what is needed if they are to grow quickly enough to survive as a company long term?

The clustered NAS vendors could cause them pain going forward, the Isilons of this world are looking to do to NetApp what NetApp did to EMC. In fact NetApp remind me alot of EMC of four or five years ago, which is ironic as much of NetApp's appeal was that they are not EMC!

And what's worse, you've got some big players who could do immense damage to NetApp; I'm thinking IBM, Oracle and HP. IBM with SONAS could hurt them at the high-end and Oracle with the 7000 series could really hurt them in what has been traditionally their heartland; medium-sized NAS environments. HP could revitalise their storage business under Dave D but that is probably a longer term turn-around as opposed to immediate threat.

Yes NetApp have spent time building some strong partnerships but even this is a bit 'meh'; not a huge amount of organisational innovation here. Nothing which made us sit up and think, 'blimey, that was a clever move!'

In fact at the moment, 'meh' pretty much sums up NetApp as a company. Lots of companies go through a 'meh' period; HDS have been sitting in theirs for some time and need to come out pretty soon. EMC went through their 'meh' moment...IBM and HP have managed to have 'meh' decades in the past! Can NetApp come out of a 'meh' moment fighting and innovating? Lets hope so!

p.s I've labelled my own post as FUD...because if a vendor had written this...I might have accused them of writing FUD!

Comments

[Sarge]

Meh...

[Dimitris Krekoukias]

I'll begin with a concession:

The absolute best way to get NetApp's value is if you migrate your data to our filers. Not use them as backup targets and the like. Which is one of the reasons NTAP wanted to buy Data Domain - several customers JUST want a backup target and are loath to make any changes to their infrastructure, no matter how much sense that would make or how much money they'd save. I fight this battle daily. But you know what? The failing economy is actually HELPING NetApp.

NetApp is not in the point solution market, which may be your entire point. But neither is it a niche provider.

There's a flipside to that: http://bit.ly/dsj1sS

Let's not forget about the rather amazing SANscreen product... multi-vendor and can help almost anyone, no need to buy any NetApp gear.

Or the Operations/Protection/Provisioning Manager suite, which is utterly amazing.

Or the application integration, that's way deeper than any other array vendor.

Now, on to your points.

ONTAP may be a single product but it has tremendous functionality built in, arguably the most of ANY storage OS out there. Which makes it VERY relevant to most. Which is the design goal.

If you look at the vast majority of potential customers for NetApp, they are perfectly served by our systems, indeed, the boxes are way overkill for many. So I don't understand where the real issue is for 95% of our customer base.

As it stands now NetApp has customers with hundreds of PB deployed, including some companies that are so high-profile that I'll wager anyone with a computer that consumes media is using NetApp. And those companies aren't going anywhere nor are they going to consume less storage. The product seems to be fine for them, and they swear by it.

There are other uses for NetApp instead of primary storage BTW:

One could use NetApp gear for archiving since SnapLock Compliance is arguably better than almost any file locking scheme out there.

Or one could use NetApp gear instead of Avamar with OSSV agents backing up remote machines. A certain large US bank is backing up over 450 sites that way.

If you look at cloud providers, many of the really big ones are already using NetApp behind the scenes, and have been for a while! Certainly NOT Atmos, which is about as baked as unmixed batter. I'm sure Atmos will get better, but it's just too new. Check this out: http://bit.ly/aXw9Yo

Storage-wise there are some niche markets where only Isilon, Panasas and IBRIX really have a fit - not NetApp, not EMC with MPFS, or BlueArc.

The scale-out NAS market is cool: "I need a single filesystem and namespace spanning 10PB! Not a conglomeration of volumes that I have to join!"

Aside from the coolness, the companies that need this kind of thing are few and far between.

In addition, the true scale-out NAS players typically are optimized for very specific, sequential I/O, and utterly suck at random I/O, making those products completely unsuitable for the vast majority of workloads. BTW - how do you back it up? And where are the other zillion features NetApp has to offer? That's right, they're not there.

So, Martin, is it your opinion that if I don't support a single 10PB filesystem my company will go under? Should I go work for Isilon? I'd find it kinda boring going to 1 meeting ever few weeks... versus the 15+ a week I average now.

NetApp has been innovating for a long time, it seems as if your NetApp team is not touching you properly, send me an email and I'll have that fixed. Otherwise you'd know what's coming and what's already arrived... :)

Even if you look at just ONTAP, the underlying code has undergone massive changes over time (the WAFL hater club seems to prefer thinking nothing has changed in 15+ years). There have been huge optimizations over time, various tunings for different workloads that you can try, and so on - things that can have a 3x performance impact, not talking 10%.

And things like PAMII are not exactly trivial and are saving customers tons of money and improving performance - not bad for tech that you can install by adding a couple of boards and an OS upgrade... other vendors haven't had something similar, it's always like "well you need the new controllers" or "you need to upgrade to the other, utterly incompatible line of our products". Or - buy SSD + unbaked software to manage data migration.

And finally - the biggest change with ONTAP8 is the large aggregates, something that only the larger customers will appreciate anyway. Yes, the product has been taking a while to become GA, but I wonder how long it would take EMC to build something like a properly unified Enginuity/DART. Marrying 100% disparate codebases and having a fast and reliable result is NOT easy! So pardon us for not wanting to release code that's not pristine.

Of course, you can always download the current state-of-the-art release and go at it, with the usual pre-GA caveats. Many large customers are already doing so.

Anyway, gotta go take a shower and go to my meetings - there are real customer problems to solve!

D

[TimC]

I guess I'm a bit confused as to what exactly it is you want out of NetApp. ONTAP is clearly going to be a slowly moving target, not night and day, flip your world upside down changes. They were the first vendor to support FCOE, and they just recently added the ability to migrate a vfiler from one cluster to another with no downtime (to NFS or iSCSI connections). There's a hundred other optimizations and changes that could be rattled off that we've seen in the last year, but the point is they're continually adding features and improving the code.

If your true primary complaint is that they didn't buy datadomain and they haven't bought anyone else... my response would be that those sorts of deals don't happen overnight. I'd imagine their executive team is looking at other targets and evaluating their books, their culture, and their products to make sure they're a fit. Personally I'm happy to see them taking their time rather than running out and grabbing the next thing they see.

As for HDS, you really lost me there. Have you looked at their product line-up at all? The AMS2000 is by FAR the best modular controller design on the market today. True active-active heads is a godsend. No FCP partner path misconfiguration errors (NetApp), no LUN trespassing (clariion), it just works. All the paths, to all the LUNs, all the time. Not to mention, first to market with a SAS backend, and mixing SAS and SATA disks in the same tray. I'd say Hitachi has been the absolute leader in the modular storage market from a pure engineering/hardware design perspective in 2009/early 2010. They don't have all the features that NetApp has (and don't get me wrong, some of those features are absolute requirements for many customers), but what they do have (fast block based storage), they're doing better than anyone else right now in the modular market.

*no, I don't work for any of the storage mfg's.

[Martin G]

And still in all of that; I just see a filer company, with the exception of SanScreen, every product you mention relies on me having your infrastructure. And I purposely avoid comparing you with EMC for a reason. I can't run my business on either of you but I could buy more from EMC without touching their storage hardware. And if for instance, V-MAX goes through a rough patch like Symm did arguably, EMC can probably pull through with CX and their supporting software whilst they sorted any hypothetical issues. If OnTap goes through a rough patch and I really hope for your sakes, that OnTap 8 is 100% rock solid; it'll hurt a lot. Because without the Filer, you've got no core product to cloth and hence bring your other products to market.

You aren't the only company in this boat by a long street but you are the most visible. And your strength is your biggest weakness.

The cluster-NAS vendors aren't going to hurt you yet but EMC didn't take you guys seriously either. Who wanted NAS? It wasn't a big market! And to your credit, you made it your own. And now everyone wants a slice. Instead of attacking EMC, I see a lot of vendors now aiming their presentations at knocking you.

And sitting above you in the food chain, you've got some big boys who are beginning to think, I want that pie as well.

Your first billion was the easiest billion; the rest gets harder...there's a target pointed in the middle of your foreheads' now. You are going to have to run even faster, even to stand still. Can you do it and can you do it without increasing your product set?

I'll give you an example, the company I work for set itself a target for a number of subscribers after which point we start to get to saturation and gaining more will be hard....equilibrium will set in.

So how do we grow revenue? We offer more products so we can sell more things, so we innovated in the world of HD; we'll have 3D this year, we bought a broadband provider...if we are going to grow, we need to do more. You will need to do more yourselves. And yes, we kill off unprofitable lines of business but we make sure we have something to replace them with.

[Chuck Hollis]

Hi Martin:

My, hasn't Dmitris turned into quite the passionate cheerleader lately?

Competitive rancor aside, I tend to think of overall business strategies, much like you do. And, you're right, NetApp has turned largely into a single product company.

None of their acquisitions have turned out particularly well. None of their efforts to move outside of their core dual-head filer business have worked out particularly well either.

As long as there's a good, profitable market for dual-head filers, they're in fine shape. Nothing wrong with doing one thing and doing it well -- many companies inside and outside tech pursue that strategy.

But, as you point out, it doesn't strike many of us as a defensible position long-term. It's just too easy for competitors to provide a "good enough" and attack those rich margins.

EMC tended to think it was in an unassailable position a decade ago with the Symmetrix. We were quite wrong. If I was doing a storage startup, it'd be pointed directly at draining NetApp's profit pool.

I also think that single-product companies lose the benefit of hybridization -- combining multiple technologies in new and interesting ways.

I also would agree with your views on "partnerships" -- NetApp's strike me as extremely assymetrical, where NetApp is doing all the work (and all the talking!), and the "partner" is enjoying the ride.

By comparison, EMC's strategy of pursuing diversification and integration has its challenges as well -- there's no magic answer to any of this.

(disclosure: I am a long-time EMC employee)

Thanks for a thoughtful post!

-- Chuck

[Martin G]

Tim,

HDS' 'meh' is their sheer inability to market their products and tell their story in a compelling way! And they also need a refresh at the high-enterprise end; they need to keep moving!

Complaint about NetApp is how vulnerable they are...they are not the only ones tho'...there are more vulnerable companies out there; 3Par, Compellant, Isilon, BlueArc etc, etc...the trailing pack who need to grow quickly or get bought. I have no problem with OnTap as a bog-standard NAS product...have a problem with how long OT8 has taken. NetApp are good for the storage market but they feel a bit complacent and almost as they expect business by right. Just like EMC did, just like IBM did, just like Oracle do, it's time to shake them a bit.

WAKE UP! THINGS ARE CHANGING!

[John H]

I always thought Dimitris worked with Netapp not for Netapp.

"my company will go under"
"So pardon us for not wanting to release code that's not pristine"

There I was thinking he was an independent ;-) Thought his recent posts had gone a bit lopsided.

[Dimitris Krekoukias]

Oh lord, "with" vs "for", there is no secret, conspiracy theorists go see my comments here http://bit.ly/asrtLf

Recoverymonkey has always been paid for by me looong before I joined NetApp - why don't you look at the pro-EMC content in there instead of trying to spread more FUD?

And why did nobody EVER question where I worked when I was posting pro-EMC stuff? Interesting, that one.

NetApp has enough corporate bloggers, and I post all kinds of other stuff in my blog - go ahead, read it. And focus on the discussion please.

Chuck, I'll take your comment as a compliment, I try to learn from the best. I always like your spin - because, ultimately, it's effective.

Back to the subject at hand:

Martin, I see you wrote "I have no problem with OnTap as a bog-standard NAS product" - thanks for that, but have you tried using the FC and iSCSI portions of it?

I agree completely that in order to increase marketshare you need to diversify.

There are more than one ways to diversify.

One way would have been for NetApp to keep the old NAS-only boxes and launch a new series of iSCSI and FC boxes, kinda like others did.

Another way was to augment the system and give it the capability of doing more stuff instead of acquiring another company or building a new product.

Look at other devices like phones (bear with me).

What devices are the most popular by far? I'll save you the trouble: multi-functional ones, like the iPhone. I'm sure I'll get grief for that comment, maybe someone will say I work for Apple.

Even EMC, who you think is unassailable due to the product breadth, could face issues if there's a huge blowup with Enginuity and FLARE both. Symm and CX account for a gigantic piece of the EMC pie, and if those go away, the company will not go away but it will be in trouble.

It's a good thing all those QA engineers work hard to prevent this from happening.

Those same QA engineers are preventing you from going ONTAP 8 just yet.

D

[ianhf]

Hmmm isn't this a hot topic - and frankly I agree with most of what Martin has written... For me NetApp 'fell asleep at the wheel' 3-4yrs ago - and are suffering from it now...

The core product is OK, and has a fairly religious following (and rightly justified in history) - but it's simply not moved on enough commercially or technically over the last few years. Yes they have some good technology (their MetroCluster is much better than Celerra for failover etc, and MultiStore better for segregation).

Trust me I'd really like them to be doing much better, but they seem so much in denial it’s hard to have a sensible conversation – others are much easier to work with and, whilst not always quite ready for prime time this afternoon, as Martin put it, “the rest of the market are catching up and some could rocket past them.”…

To comment on some of Dimitris' passionate points above :-

Yes SANScreen was indeed a great product (much much better than anything EMC / Ionix or anybody else have released), but the spark is rapidly vanishing... Feeling that the innovation & enthusiasm has been snuffed :(

Hard to understand how the failing economy is helping NetApp, their storage is among the most expensive out there, with the least flexible procurement options, and the most disinterested account teams – taking the most effort of anybody on the customer’s side to get competitive pricing

Application integration - tricky one, especially when a large part of the requirements for enterprise storage are moving to be commodity interfaces that specifically don't want vendor lock-in (and charged for) integration technologies

Large aggregates 'only for larger customers' - hmmm wide-striping within an entire array is for all... Oh and as far as I'm aware I can't move existing aggregates to new 64bit sizing without an offline migration of data...

Re "So pardon us for not wanting to release code that's not pristine." - does that mean 8.0 will be released as GD immediately then?

Some more of my points :-

The boxes :-
* Not enough network interface points – we need more interfaces not more bandwidth
* Where's the compress functionality?
* Where the tiered disk storage ability?
* What is happening for million+ file NDMP backups?
* What is happening for for object interfaces?
* Large wide-stripe sets?
* Power-down disks?
* 'Recommended' use of large SATA disks? (NetApp engineers don't appear recommend using > 750GB SATA due to aggregate performance issues)
* What is timeframe for the GD (not GA) release of OnTap 8?
* When will the non-disruptive upgrade to 64bit aggregates be available in GD release?
* When will we be able to use 250TB+ arrays without sacrificing the majority of the interface ports?

NetApp had a strong & good relationship with Oracle for NFS database storage (much much better than EMC), but it’s very hard to believe this will continue to near the same level given Oracle now have all the storage technology they need and that integrated database appliances (al la Exadata v2) will be common-place in data-centres shortly.

I do see all of the ‘established’ NAS providers being impacted by the server manufacturers, with it being possible to ‘roll your own’ more easily & low cost for a certain segment of the market.

And of course, when are NetApp going to release a TCO/ROI model for their technologies and when will they realise that in order to compete their prices need to change? In short “wake up, smell the coffee and do something big & quick!”

For me NetApp are the company I’d love to like more (as I did in the past) and would love to have better evidence to argue their case and do more business, but sometimes it’s just such hard work for that you really do think ‘meh’…

[EricH]

I see a lot of feedback here from a very narrow set of individual customers. Each customer is of course always right, but my observation of NetApp is that they are chasing whales (i.e. entire markets) instead of merely fishing for specific customers.

My work with a GSI recently opened my eyes to the enormous "once in a lifetime" market opportunity for truly virtualized infrastructures at all layers of the IT stack.

NetApp seem extremely well positioned (both technically and operationally) to capitalize on this trend with their existing technology. The market clearly bears this out given yesterday's stunning results.

Obviously not every customer will be pleased with NetApp's roadmap, but it seems entire markets will - and that's why I would never sell NetApp short (literally or figuratively) in this market.