The current turmoil caused by plummeting cloud storage costs, new entrants sporting modern architectures and the forced re-architecting due to flash and upcoming NV memories is a perfect storm for legacy vendors. Some are handling it better than others, but some, like IBM and NetApp, appear to be sinking.
NetApp is signalling that their 2015 sales may not reach expected levels – not a surprise – and that more layoffs – on top of the ≈1500 in the last couple of years – could come soon.
The latest troubling sign from NetApp is the failure of their FlashRay project to ship a competitive product. The VP in charge, Brian Pawlowski, left NetApp for Pure Storagelast month and the company folded the development effort into the ONTAP team.
According to press reports, the FlashRay project started almost 3 years ago, but has yet to ship a competitive product, despite the efforts of 100+ engineers. Given that there’s nothing FlashRay was supposed to do that was terribly novel, the problems are likely political than technical, a conclusion reinforced by placing FlashRay under the ONTAP team.
NetApp has well-known problems integrating acquisitions into their products. Now it seems they have problems developing new products as well. Not promising, given the threatening secular trends pinching their core business.
The secular trends include:
File server obsolescence. NetApp’s original raison d’etre is no longer state of the art – or all that interesting. File servers were a great idea 25 years – like RAID – but newer technology – object storage – is looking to replace them.
Cloud encroachment. When arrays were the only game in town, customers had to buy another when they needed capacity. But now old and rarely used files are moving to the cloud – and the repeat business with them.
Margin pressure. The combination of object and cloud storage – both much less costly than traditional arrays – is waking up customers and producing buying resistance.
A recent note from a disgruntled NetApp customer highlights this issue [edited for length]:
I’m currently engaged in replacing a FAS3240 going off maintenance with a newer version because Netapp offered a deal I couldn’t refuse. . . . [W]hile the drives and shelves are at least not inflated to ridiculous multiples (Why the heck am I being charged even a penny over retail?) they’ve now instituted a $34/TB surcharge to “license” me to use the very storage hardware I just bought from them.
. . . I am incensed at their audacity to charge me a FAT 60-70% margin on commodity hardware. . . .
The reason that customer feels he is being gouged is that the direct sales model is expensive. You pay for your salesman whether you like him or not – and he’s paid a lot more than your local Best Buy clerk.
The StorageMojo take
Bulk file storage hardware is commoditized. The upstart vendors are virtually all software-only, offering tin-wrapped software through resellers rather than direct sales, which saves a lot of margin dollars.
NetApp is in the position that DEC was in the ’90s, where commodity servers eviscerated the VAX business, aided by supply chain innovations by Dell and others.
Servers became a box to be ordered over the phone, not painstakingly configured with a sales engineer and delivered in a few months. Storage is finally following suit. And it doesn’t look good for NetApp.
Courteous comments welcome, of course.
Agree that the storage platform/solutions landscape is evolving across the board. Not just for File Managment NAS platforms, but for all HW storage platforms. However, I don’t see ‘traditional external storage’ going completely away anytime soon.. All this new ‘disruptive’ storage technology and software defined platforms aren’t the only driver. “The Cloud” in whatever form; Virtualization and things like Data Lakes, Hadoop, Docker/Containers, Object storage; Mobile Computing, IoT, etc. are major transformers too. None are mutually exclusive or can totally eliminate HW platforms. BTW…your DEC commment was somewhat accurate…but of course there’s more to THAT story. I was there…for years. We were in the commodity IA and RISC based server market too by the mid 90s. “Open” standards and market momentum favored MS and Linux SFW. VMS O/S should have been opened, but Ken Olsen opposed that until it was too late.
How does this trend only affect NetApp? Don’t the same issues affect EMC, HDS, XIO, Nimble, Tintri, Tegile, and the storage businesses of HP, IBM, and Dell?
Seriously , “why the heck am I being charged even a penny over retail?”
Buying hard drives on their own does not = Enterprise class storage. I work for NetApp and I have never been so excited about the future, we are filing more patents than ever. Our Flash Systems are constantly beating “built from the ground up” startups in POC’s. We still over the best application integration in the industry with ONTAP.
Add to this the great acquisition that is SteelStore and the rest of our portfolio. Maybe we do a bad job at responding to FUD and we need to better at marketecture. We have the best storage O/S there is , we have startups going into clients pitching technologies we’ve had since 1992 but with no unique IP.
Netapp provides good business outcomes, sure we could do some things better but we have over 2,000 developers working on ONTAP. Code changes, we’re a smart company with smart people.
How doomed are NetApp? I’d back this horse over any other.
We were Netapp customers until two years ago when we purchased a Pure Storage array. We used the Netapp as a NAS box for file/folder storage and NFS for vSphere. A little nervous going to a straight SAN but after we front ended our file shares with two virtualized Windows file servers we found them much more flexible than the Netapp NAS capabilities. Couple that with the dead easy interface and great performance of the Pure box and we have never looked back.
NetApp should’ve done a long time ago what Nutanix did, instead of doing a toxic partnership with Cisco and get involved with the Flexpod.
If NetApp is going to acquire anyone, it should be Nimble. A large % of their staff is from Netapp, and understands the model and faults, which would help them in the long run. Not so much Pure Storage, as it is doing OK in the enterprise level and has a lot of investment from big investors not only on the capital raising side but actually using it in house. (Bridgewater Associates being one who just replaced their entire VMAX environment with Pure.)
> we are filing more patents than ever
This points to a future as a patent trolling^Wlicensing entity, I’m afraid.
ONtap the best software, give me a break. It took you 5 years to get clustering to work and it’s still broken. You guys are so far behind now compared to 2010. I don’t see any way you don’t get bought for parts in the next 12 months.
What about DDN’s WOS? That looks to be the future.
> Buying hard drives on their own does not = Enterprise class storage
True, but you (netapp in particular but 3par, IBM, EMC, et. al.) have some SERIOUS delusions of grandeur! I run a modest FAS3240 system. After 3 years you demanded an obscene figure for extending support. While I understand you want to get people off old equipment, one rev behind hardly qualifies. I could forgive a 10-15% surcharge over the previous term, but not 150%. And your reps/resellers in NoVA at least, refused to bless a configuration that retained my 600G/15K drives when you don’t even supply that speed with the replacement 80XX family.
But the icing on the cake was not only charging at least 20% more for the commodity hard drive (which you guys buy by the container load) but then you think it’s proper to “license” me the very capacity I just bought. You pull the same nonsense with the E-series; just at 60% of the FAS markup.
I heard all kind of moaning from the pricing rep (a Netapp employee) about how this system I was buying was worth 800K and that I was getting such a sacrificial deal at 70% off because it was end of quarter. Excuse me? The solution priced out at $300K was still a healthy 50K more than it had any business being. I’ve gone round and round with Netapp over the years and this was the last straw. Your platinum dealer in northern VA is now on my permanent “do not do business with again” list. I went across the country and that reseller didn’t even try and cut 20% off the tab. You can forget ever selling me another system. I will gladly build something from white-box or engage a vendor who actually wants my business and isn’t so full of themselves WRT their software value position.
Maybe in 1980 unlimited snapshots, and “zero” impact Raid6 writes and such was a unique differentiator. It hasn’t been for neigh 10 years. So charging 2x what your competition does for the same feature set says, Netapp wants to lose business.
If Netapp is to survive at all it is as a software company and chopping their bloated software headcount in half. Get out of the hardware business. All of it is OEM’d from Xyratex anyhow. Nobody believes the story (from EMC, Netapp, IBM) about doing any hardware engineering.
Netapp SE.. why even comment? Such fluff and Cool Aid.
When I asked out Netapp account manager what their strategy was for SDS, and Cloud, his response was they plan to ignore it and assume all the customers they lose will come back. SDS wouldn’t prove to be reliable (QA required to support various HW/OS combinations), and the dissatisfaction with the cloud (short term thinking, performance, security). I went to their recent executive road show and many of their talks seems to support this. They had their in house council there and spent an hour on data sovereignty. It was a good topic but…
How many companies ignore competition and survive?
“I have never been so excited about the future, we are filing more patents than ever.”
s i g h
Great article, but the one thing that seems to be missing from NetApp unlike say EMC is a large s/w division selling “storage” related solutions – such as backup and archiving. Its clear that even NetApp’s Enterprise s/w is pretty poor. – thats why they have to rely on OEMs with other companies.
Isn’t Netapp’s snapshot app integration just comm-vault OEM code? (Serious question?).
C-Mode’s lack of aggregate wide striping makes me feel that they shouldn’t be allowed to call it scale out.
The crazy pricing for years 4,5,6 of support is the leading cause of clients leaving FAS in my territory.
Windows 2012R2 is actually a damn good NAS OS for SMB3 (better native reporting with FSRM, than a stock FAS). If I want NFS at large scale beyond XFS and RedHat I honestly prefer HNAS, or one of the niche specialty HPC or true scale out vendors.
First of all I am a former NetApp employee who worked several years in the support center and then the professional service group.
The NetApp SE’s comment is clearly a biased glaze on a pile of poo which is to be expected because an SE is a Systems Engineer = Sales Engineer. They are technical engineers who are often times pretty good at what they do to include generate revenue for the Sales team because SE’s get commission as well.
In terms of NetApp’s technology, it is no surprise they continue to deliver crappy products because there is consistently high turnover rate in the development teams as well as the technical support center. The TSE’s are well aware of technical & political issues that plague the company and day-in / day-out they try to do the best for the customer but management does not give them the tools to do their job effectively. As a matter of fact management practically encourages the TSE to lie to the customer and get them off of the call because they are extremely short on manpower and need to get calls rolled over. One of the comments above mentions about Nutanix having many ex-NetApp employees which is an accurate statement. Most of those engineers love what they did at NetApp and were always for the customer but the company tied their hands which ultimately caused a mass exodus of which many ended at Nimble or Nutanix. About 90% of the top talent in the NetApp Global Support Center in RTP left because of the politics first and customer last mentality.
The clusternode technology that was not ready for primetime in OnTap 8 was actually a technology (GSX) purchased from another company. Bugs often times go for several quarters before they are even reviewed but falsely conveyed to the customer that a fix is in the process. What’s an even bigger joke is the hardware specs of the FAS32xx / FAS8xxx is you can essentially build your own server at a tiny fraction of the cost. Get infiniband adapters and cables on eBay and you’ll have a 10 node cluster easily for much less than half the price of a FASxxxx from NetApp. What you’re paying for is the 6 figure licensing per node which is essentially disgusting. Lastly, the deduplication which was one of my specialties was a massive failure when it comes to databases and mail systems.
Legacy storage is doomed: Expect converged, or hyperconverged, solutions to reinvent the industry … quickly.
Kenneth, there will always be a place for modular storage (PLENTY of environments have data that grows asymmetrically to compute). Hyperconverged is fine if compute growth is linear to storage, or keeps pace with storage, not the other way around. (I run VSAN for my production cluster, so Its not that I don’t love the concept/tech, just there will always be deep archive etc that it doesn’t work for).
Will I’m not quite as doom and gloom as you on the details, but more “for the premium” I don’t see how they are keeping pace with the other big storage Tier 1’s EMC and HDS (Who are moving farther into data management, and sexier, valuable pieces). Or if you just want a bunch of storage someone like HP who can ship kit for cheaper.
You have to defend your margins with ever refresh, and that is the tough part. Having seen people deploy a new 6-8K ever 12TB because of single thread CPU bugs, I’ve seen some downright Stockholm syndrome Netapp customers, but I am seeing people who are having problems moving on. This is the part they need to “wake up” and take more seriously.
For what its worth (Audience), who else has a good pNFS solution right now?
> who else has a good pNFS solution right now?
I’m afraid the answer to that is nobody. Ok, maybe FreeBSD has finally got that sorted(?). For all it’s problems, Netapp does NFS, the best.
NetApp is dying. Tom G is the worst CEO I ever see in this company. The politics he brought into this company is killing this company. The biggest mistake NetApp executions made in these years is acquisition. No matter Extreme I/O, Pure Storage, or Nimble, buy them and throw dammed money on their faces, and occupied the Flash market and then integrate with Ontap step by step. NetApp’s acquisition is never successful. they failed from data domain to the latest flash competitors. The 2nd mistake they made is lack of execution. From integrating cluster ONTAP to Flashray, everyone got used to delaying. The execution will move more jobs to india by cutting more jobs from SVL, RTP, and etc. It will further decrease the innovation of this company. NetApp has been used to mediocre. They still consumed their last century’s legacy: WAFL and HA pair without a little innovations. It’s not easy to turn around this company: their executions still don;t realize the root cause of their problem. They are lack of a true leader like jeff bozes, with real vision and understanding of technology. Keeping buying back stocks will not save this company.
In general, the golden age of enterprise storage giants has been gone. It’s impossible to still get so high margin with just giving out software license, like IBM, EMC, NetApp. The cloud storage like AWS, Azure, Google storage platform will dominate this market. They have advantages the traditional enterprise companies don’t have: 1) cheap and on-demand charge. 2) provide free and reliable softwares (cloud interfaces) to process data. 3) object-oriented. And data storage price is really reflected into their lifecycle: from cold to hot, from cold storage(AWS Glacier) to on-line data (S3).
To NetApp SE, I too am a NetApp SE, but you have to listen to your customers. Our customers are telling us we need to change things, that we are behind, overpriced, and not competitive. We can believe how great we are and spout all the internal NetApp marketing jargon, but until we start listening to our customers we have nowhere to go but downhill. I worked for well over a decade at another storage company and trust me, those within NetApp have a very strange idea of reality. Just like NetApp has some really great features that others don’t, our competitors have had features for 10 years that we still don’t have.
A lot of NetApp hating going around. I’m an ex-NetApp employee as well and i’ll just say this:
Our 64 NetApp clusters are managed by two people. These clusters runs everything but iSCSI and house all of our NAS.
Our 12 VSP frames have no less than 8 on shore engineers that support them. They only do FC.
HNAS better than ontap makes me laugh. It’s the most half baked crap NAS solution I’ve ever had to use.
In an environment with three fully redundant datacenters NetApp has the features we need and the ease of use to make what takes 8 people to manage from another vendor only take 2.
All the startups look awesome, and they are, but while I agree NetApp is plagued with horrible leaders and decision makes over the past 5 years I think they have plenty of time to get their shit together and still be a market leader.
Hey Netapp SE,
Someone forgot to tell your sales people about your enthusiam. Netapp’s revenue has been flat for the last three years and net income was higher in 2011 than it was last year. I worked for Netapp from in the 90’s well into the 2000’s when the only pull backs were due to macro economic events (dot com bust, great recession) 2010-2014 wasn’t a period marked by such an event but the company appears to have hit a wall by 2012. I’m reminded of a story Warmenhoven recounted about a meeting years with Steve Kleiman who once asked him if Netapp wasn’t just a disk. (ie. Are we just a software company?) It will be an interesting road ahead for Netapp.
Ex-NetApp software development engineer here – worked for 7 years and filed 4 patents. Let me assure you, these patents were all crap. There was nothing innovative about them. Any reasonably smart software engineer can come up with these ideas. Top-level management encouraged filing patents just to save their asses if anyone were to sue them in future – it’s normal practice for any big software company – and engineers loved submitting patents because it brought them extra cash.
Dan Warmenhoven was the right leader for NetApp. It was pretty much downhill after he stepped down. Nimble kept recruiting all smart engineers from NetApp, but management did nothing to squish and kill the little bug which it was back then. I hated NetApp for that – it was like they didn’t care! I couldn’t believe it kept getting ranked the best place to work for. There is no diversity (just too many Indians), perks are mediocre, and communication is not open.
Manish Goel and Beepy leaving seems like the final nail in the coffin.
I spent almost 9 years at NetApp during the early/mid 2000’s and it was a wonderful working experience. Great executive leadership, highly disruptive technology and a keen customer focus. But times have changed; for the worse.
NetApp’s circa $300M acquisition of Spinnaker Networks in 2004 and the disastrous circa 10 year/multi-billion dollar ($$ are a WAG) integration to deliver a solution that has completely missed the massive Flash disruption is akin to IBM missing the PC revolution. Unlike IBM though, NetApp’s PS/Outsourcing business isn’t going to drag them out of the hole, and hiding behind a “Cloud” strategy isn’t going to work either. At the end of the day, complexity hidden by a veiled curtain called “Cloud” still translates to a per unit cost at the pointy end. Make no mistake, NetApp is a very complex solution today, just look at the PS numbers.
NetApp has roughly three options available:
1) Do nothing and death spiral results,
2) Get bought for the cash in the bank and a little upside, or
3) Buy a Flash based solution that is accretive to the existing business, and that is not Nimble or Pure/SolidFire or Nutanix/Simplivity, but could be Tintri.
Unfortunately NetApp’s track record for successful product acquisition is Zero, well maybe there is a .25 each for Akorri Networks and Engenio, but that is only slightly above their ability to develop and bring a great new product to market since the R100 (Disk based backup) in circa 2002.
Option 1 may have the inside running.
“How doomed are NetApp? I’d back this horse over any other.”
Ive met this guy, seriously..I -have-. Let me describe him.
He was the last guy at SGI.
He was the last guy at Sun.
I recall he also used to talk about a roomate at Webvan..
And his dad was the last guy at Shockley Semiconductor screaming “Go ahead and quit..I’d back this horse over any other.”
I am an ex (as of Tuesday) NetApp employee and think the layoff was a gift. I was looking around for a new job and now I get a vacation before I take one.
I worked supporting sales and internal employee development. Over the last 6 months I have been asking around about what people thought about the challenge presented by Ceph, Hadoop, r-ram and so on and most people had not even heard of them. I assume there is some part of the company that was thinking about it but my job brought me in contact with many parts of the company including engineering, low level management, sales and support. Sales knew about it because it was killing them but no one else seemed to. The few people who had would say “enterprise customers want the support of a trusted organization so they are willing to pay the extra.” That sounded just like the death rattle of both DEC and Sun.
That scared the s*it out of me and that’s why I was about to bail.
Also, to what ShutApp said above. The manager of the group I was in was Indian and no Indians were laid off.
@Netapp SE, “Buying hard drives on their own does not = Enterprise class storage” why not just unlock the commodity harddisks support. I doubt if any storage company would have done significant firmware changes to the OEM drives.
I’m a new hire to Netapp the day they announced they were laying folks off. I heard the news and it did not phase me. It was expected. Shocked they did not lay off more management. Current management appears to be the problem. An all hands resulted in very little questions from staff. Leadership is clearly a problem. It’s not going to get any better….
I was laid off by Netapp 3 years ago. I was at 154% of Target, yet selected I feel, because as a Cloud Architect, I was warning management that Storage was a commodity, and they needed to sell BULK storage to ISPs, not complicated, old school model of trays and disks. I also told them that Storage needs to be priced like IaaS or SaaS, by the drink, not the truckload.
ISPs like Comcast, Verizon, AT&T and Savvis were looking to design and build offers based on compute and storage as a service, but NetApp wanted to sell by the BOX, not by the workload, capacity, and throughput.
Its the abstraction model coming back around. Mainframes presented resources for users by TIMESHARE, small servers and workstations broke that down and crushed the costs associated with mainframes. XEN and virtualization reintroduced abstraction, then clusters of VMs and Clusters of Data Centers abstracted things once again, crushing the costs of components. Now Clouds, and MetaClouds or Clusters of Clouds are again abstracting the industry, and crushing component costs yet again, and if providers don’t float up higher and higher on the wave, they will lose pricing ability even further. NetApp needed to design, build and put it’s money where it’s mouth was, and own and operate a CLOUD of storage for others, not keep trying to sell it’s storage to buyers who are leaving in droves for lower cost options.
If NetApp’s products are so great, and it’s cloud offerings so wonderful, why don’t they drink their own koolaide, and build the most cost effective, automated storage cloud on the market? Then sell the storage service… Beats me, but I proposed that 3 years ago, guess I got my just deserts…
You don’t need a crystal ball to see NetApp’s future as a patent troll or portfolio for sale. Don’t forget NetApp has already exhibited patent troll tendencies in the past when under threat from cheaper solutions. It sued Sun over ZFS then attacked smaller companies that used ZFS and couldn’t afford to fight back. http://www.theregister.co.uk/2010/09/09/oracle_netapp_zfs_dismiss/
It’s actually amazing that you are even reading this. Haven’t you read enough? Well my side of the story is NetApp is crap. The support is crap and its amazingly insanely expensive. Don’t ever buy one.
How doomed is EMC now? -oh ya, EMC just sold themselves out to Dell.
I’ll stick with my ever expandable and more adaptable NetAPPs.
Need to update this. Netapp has Solidfire, Malinia, Flexpod for openstack. NetApp is in play. NetApp as a company contributes more to the OPENSTACK community than anyone else.
i have netapp storagevault every part of it requites licssens.
u look at cost of unit, and each liccenes u need. dont make sence
u by a ibm 3850 m2 put a sas array on it and load up with seagate 2tb es harddrives will set u back under 400.00 for all os only thing needed
the power of this server speaks for its self as ibm still supports it with drivers. at up to 24 cores how can any nas compare in fack u can install nas kind of os on it. netapp has one huge probem that is they dont encript the drives or os, my stoagevult has 10 500 gig drives and for most part configer and i so hate it. the http config opton not there all files are missing iscsi only works in part on 2012 domain wont work on windows 7 i cant even log in unit from windows 7 i think might be do to the encriton the network is set up to use but why would not netapp be conpattable with it,
Curious about your take two years later.
LOL NetApp doing all right now.