NetApp Finally Pivots

Last month I attended Storage Field Day 9, which included a two-hour presentation at NetApp headquarters. My summary statement of what I learned from the briefing is that NetApp is pivoting. By that I mean that they seem to have finally decided that OnTap is not the answer to every problem. I wrote about this a little in my post about AltaVault and OnCommand Cloud Manager.

The main criticism of NetApp over the years was that they were a one-trick pony. Dave Hitz talks about this in his book, How to Castrate A Bull, which is his story of how NetApp was created and shaped over time. Calling them a one-trick pony is quite a stretch because their products do so many things beyond what the original NetApp filer was originally designed to do. But in the end, it still all came down to Data OnTap, their core operating system. Many argued that their complete reliance on Data OnTap is what caused them to take so long to leverage the IP they got from the Spinnaker acquisition. They could have had a clustered operating system so much sooner if they didn’t require it to be built on Data OnTap. But now, of course, they have Clustered Data OnTap, or cDOT.

But in the last year or so, they’ve made a few moves that signal that they realize that they must be all things to all people in order to be able to sell something to everybody. Their products must be cloud friendly and they must have a proper flash story to play in today’s marketplace.

For a long while, they were happy to sell you an all-flash array based on cDOT, but someone inside NetApp recognized the wisdom of an architecture that was built from the ground up to take advantage of flash — and they acquired SolidFire. Now they have a completely separate product that they can sell to those looking for an all-flash approach, and it appears they’re going to keep it that way.

The acquisition of SteelStore from RiverBed (now called AltaVault) is another interesting move. It closes a hole in their portfolio created when they discontinued the NearStore VTL product in 2008. They hoped to fill that hole with Data Domain in 2009, but that didn’t work out the way they planned.

While the AltaVault product does not have either the mindshare or market share that Data Domain has, it does come with some features that Data Domain does not have. Specifically, it can replicate its backups into over 15 public cloud providers, including Amazon S3 & Glacier, IBM Softlayer Object Storage, and Microsoft Azure. It can also replicate into several private cloud products, including NetApp StorageGrid, EMC Atmos, and OpenStack Swift.

They’ve also added the ability to accept replicated backups from cDOT, in addition to already being able to accept backups from several backup products. This means that they have a product that can accept backups from just about anything and replicate them into just about anything. Now they’re developing an orchestration product to manage the whole process, called OnCommand.

StorageSwiss Take

Time will tell how much of a pivot this is for NetApp. But it is interesting to see them taking two major steps toward the future: purpose-built flash storage and migrating data between clouds. And if the shipping version of OnCommand has half the features that they talked about, it will offer NetApp customers a number of options for moving data into different cloud providers.

W. Curtis Preston (aka Mr. Backup) is an expert in backup & recovery systems; a space he has been working in since 1993. He has written three books on the subject, Backup & Recovery, Using SANs and NAS, and Unix Backup & Recovery. Mr. Preston is a writer and has spoken at hundreds of seminars and conferences around the world. Preston’s mission is to arm today’s IT managers with truly unbiased information about today’s storage industry and its products.

Tagged with: , , , , ,
Posted in Blog
19 comments on “NetApp Finally Pivots
  1. Tim Wessels says:

    Well, there is only so much you can do with a decades old architecture and the software that runs it. NetApp is an example that nothing fails like success. With all of their in-house engineering and product talent, why did NetApp have to pay $870 in cash for SolidFire? The answer could be that after a number of losing quarters, and a significant decline in the value of NetApp shares, the company was forced to bust a move to try and stop the company’s slide to oblivion. Personally, I think they are getting the company ready for sale, witness the dismissal of 10% of their global employee headcount earlier this year.

  2. wcurtispreston says:

    I think that’s a LITTLE harsh. An architecture is not bad because it’s decades old. If that were the case, SCSI, NFS, and SMB would be dead. Instead, they continue to be the main way files and blocks are moved around, even though they all came out in 1984.

    Yes, they needed a marketing boost, and I think they saw this as a good way to do that. They’re not abandoning WAFL. They’re just admitting that it is no longer the one size that fits all.

    And FWIW, layoffs don’t mean you’re getting ready to sell. It just means you’re getting lean. We’ll see how your prediction goes.

    • Tim Wessels says:

      Well, I’ll stipulate to my comment as being harsh, but a storage architecture that has had more face-lifts than Phyllis Diller is still old. SCSI, NFS and SMB are storage protocols not architectures. Yes, there hasn’t been much new in storage for the past 30 years, but serious storage changes are upon us now with flash, in-line memory and object storage.

      NetApp needs “something” after a string of losing quarters and a serious decline in the value of its shares. Not sure that a “marketing boost” would satisfy NetApp shareholders as much as selling the company.

      Yes, there are some metrics that businesses use to keep employee headcount at a reasonable level, like annual sales per employee. That said, EMC is dumping 6000 employees as part of its acquisition by Dell.

      Yes, we will see how my prediction goes. When NetApp hits the bargain basement, I’ll predict Oracle will go shopping. Oracle loves a bargain.

  3. ericmb says:

    I ve heard for more than 15 years that Oracle were going to buy NetApp. Good luck with that theory. Then again I guess the more its repeated the more the chance of it happening ey..

  4. Eric Slater says:

    Horseradish… 🙂

    Anyone thinking a company with 5bln revenue, 90% of which is tied to ONTAP will wake up one morning and say “forget” I’m going to “pivot” is mistaken.

    Netapp may seem to be changing to the outside world but they’re masters at saying what people want to hear because arguing which they did the last 5 years obviously didn’t get them anywhere.

    Therefore, why continue to fight market perception when you can just act humble and say “yeah, we’ve made mistakes but we’ve learned from those and we’re changing”. Fighting perception is never a winning strategy…

    All Netapp is doing is playing a waiting game attempting to weather the storm, counting beans. ONTAP IS IT. You may disagree but Just be patient…

    Has anyone looked at what ex-Netapp employees are saying? Places like glassdoor.com? thelayoff.com? Has anyone talked to engineers who have recently departed on what they think?

    This is a stubborn company that tends to admire itself for no particular reason the last 5-6 years. A TOTAL mess with mediocre executive management.

    Within 1-2 years either Netapp gets acquired or goes private. Oracle will not acquire it because they are already in the same accounts. One reason to make a multi-billion dollar acq is for account access. Oracle doesn’t need that and they certainly don’t need to acquire a declining, highly complex architecture people are staying away from.

    SolidFire? All I have to say is Meh…but the check has cleared for sure…

    Sorry for being harsh but I’ve had enough of the Netapp spin.

  5. wcurtispreston says:

    That’s why they call it an opinion. 😉

    I’ve known the NetApp story for a long, long time. And this is the first I’ve seen them acknowledge that OnTap is not the solution to all problems. If that’s not a pivot, I don’t know what is.

    As to the glassdoor/thelayoff.com… not really to the point of my post. Just said that it looks like they’re trying to pivot. We’ll see if they’re successful.

  6. Jim Weingarten says:

    NetApp is facing even more critical head winds than the transition from spinning disk to flash storage. The commoditization of storage capacity and the move toward software-defined infrastructure are the real dangers to NetApp (and all enterprise storage vendors, for that matter). As the capacity of flash devices hits petabytes (3D NAND anyone?), the cost per GB of storage will drop precipitously. At that point, the data protection, storage efficiency, and performance features that ONTAP developed to solve the limitations of spinning media will no longer be compelling. At those higher capacity levels and lower price points, maintaining redundant copies of data across clustered media becomes cost-efficient; Not to mention the power and cooling savings garnered by the elimination of all those motors driving spinning platters. Now add the move toward cloud storage as a service and the future gets even murkier. The final nail in the coffin will be proliferation of “Memorage” (persistent RAM), blurring the lines between storage and working memory. NetApp may be pivoting, and SolidFire is definitely great technology, but the days of 60% margin on enterprise storage are fast waning. I have friends at NetApp and I wish them the best, but I see a long, slow bleed for NetApp in the future followed by an acquisition at a shadow of its peak value.

  7. Tim Wessels says:

    Well, I agree and shareholders are unlikely to look forward to a “slow bleed” further eroding the value of their shares. A sale sooner rather than later would be more attractive for NetApp shareholders. Elliott Management, a hedge fund always on the prowl to increase shareholder value in undervalued companies, sold off its 10 percent ownership of NetApp shares a year to two ago. So other than its acquisition of SolidFire, there may not be a whole lot of IP there that would allow NetApp to get “top dollar” for the company.

    • ericmb says:

      A few comments.

      1. “As the capacity of flash devices hits petabytes (3D NAND anyone?), the cost per GB of storage will drop precipitously. At that point, the data protection, storage efficiency, and performance features that ONTAP developed to solve the limitations of spinning media will no longer be compelling”

      Q: This is not exclusive to NetApp? Surely it applies for all vendors?

      2. “Now add the move toward cloud storage as a service and the future gets even murkier”

      Can you elaborate in what products you think NetApp are lacking for a fit to the cloud?
      I know what NetApp Fabric is/can do and I think you ve missed the boat there. Surely being able to move customer transparently between cloud providers is awesome for customers?

      3. “The final nail in the coffin will be proliferation of “Memorage” (persistent RAM)”

      Again, why is this exclusive to NetApp? Again, this is a wide open statement applied to 1 vendor whilst it will apply to all.

      • Jim Weingarten says:

        I did not suggest that these problems were exclusively NetApp’s. In fact, I suggested that the issues will affect all enterprise storage vendors. As for NetApp’s ability to migrate workloads to the cloud, seamlessly, the cost models don’t line up. Software-defined storage has much lower capital requirement; particularly when it can be implemented using “pay as you go” and expanded on the fly, non-disruptively. The historical management benefits NetApp offered have eroded as the market has evolved (and we can certainly discuss the disarray in NetApp’s management tool portfolio). Also, there are too many vendors that offer inline dedupe, compression and replication as a software feature set on generic hardware. Move workloads into a container model, and migration is both easier and storage agnostic. The poor execution of the cDOT transition and the painful, disruptive reality of the cDOT migration opened the door to competition in NetApp accounts and drove its channel partners to look for more compelling and profitable relationships. Many of these issues are not NetApp’s alone, by any stretch. Unfortunately, NetApp has been slow to place bets and adapt. I believe NetApp’s inertia and focus on cDOT as its future are finally catching up with it.

      • ericmb says:

        Hi Jim,

        Thanks for your time and answer. I wouldnt venture into discussing cost models etc. if only because often cost models are superseeded by sales models: ie. pre sales techies getting bonuses for selling a preferred storage vendor whether its a good fit or not. This is common and somewhat obscures cost models.

        When it comes to place bets and adapt: what are your thoughts around being a customer with say 1000 VMs on cloud provider X. This cloud provider changes its T&Cs and puts its prices up. If you at this stage reside on NetApp storage you can migrate over to new cloud provider Y without disruption. Does providing this kind of feature demonstrate strategy and adaption to the cloud in any way shape or form?

        Lets not forget that ONTAP has had virtualization features for more than a decade already. This is truly visionary. cDOT is no less.

        Cheers,
        Eric

  8. Jim Weingarten says:

    Eric,

    All the vendor spiffs in the world won’t stop the shift from traditional, enterprise storage to generic, modular, scalable hardware with software intelligence. The smart vendors will go where the heat is. The channel sells what the customer wants. Adapt or die.

    In the end it’s about what I refer to as “operationalization”. Does a storage platform (or any infrastructure for that matter) offer adequate functionality at a lower cost of acquisition and operation? Can the customer access the full value of the platform through its APIs? Does the solution address the key use cases at a lower cost with comparable performance and reliability?

    NetApp had an advantage and, in my opinion, they squandered it. ONTAP and WAFL were designed to address issues related to the performance and reliability of spinning disks. With the advent of Flash, those problems changed. NetApp’s monomaniacal commitment to ONTAP, at the expense of real R&D (setting aside the FlashRay fiasco) was a reflection of bad management and a lack of vision. Competitors with more agility and less historical baggage have caught and passed NetApp. The horse is out of the barn.

    I am very familiar with the “near-cloud”, direct connect storage solution you refer to. Despite years of NetApp pushing it at places like Equinix, very few customers have really implemented it. Other more compelling solutions are emerging. A multi-cloud, workload-brokering ecosystem is already growing. At best, the NetApp solution is a transitional approach, in my opinion.

    Virtualization is also changing. VMs enabled workload mobility and vastly improved efficiency. However, VMs are no longer adequately granular, as customers are demanding greater operational agility and tuning capabilities. Containers will never fully replace VMs, but they are changing customer expectations. NetApp traded on its virtualization leadership for years. VMware never really embraced NFS, and attempted to subvert NetApp’s differentiation through initiatives like VAAI. Now VMware is a direct competitor, positioning VSAN against NetApp.

    cDOT was a great idea but it took seven+ years to reconcile the code bases and get it out the door. Even then, cDOT migration is very painful for NetApp customers. cDOT requires complete admin retraining, and there is (still?) no path to move to it that does not result in a disruptive upgrade process, possibly including swing gear. If customers and resellers have to experience that kind of pain, they are likely to consider other, more compelling options.

    Best,
    Jim

    • ericmb says:

      hi Jim,

      FYI, I work in storage OPS, I live this every day, so I have to challenge you on a couple of things,

      – VMWare and NFS are friends, best friends as a matter of fact.
      – VAAI has got nothing to do with ‘opposing’ storage, its integrating with storage.
      – Finally VSAN is not specifically against any storage vendors, its VMWare wanting to sit in the
      middle and be everything for all, all storage. Its vendor agnostic.
      – There is support for APIs etc.

      Yet for the most part I agree with your sentiments, there are massive changes occurring.
      I feel like the storage industry maybe going through what SGI went through 15 years ago when cheap linux stations took their market away: commodity hardware and software combined together took over. Whilst SGI is not at its former heights its still there with solid products. Storage products too.

      Its useful to know that NetApp release its SDS product, ONTAP edge, in 2012. It can be run on any hardware, commodity or not. With that in mind once again NetApp have been visionary and have already adapted. It supports containers, has done for a while. What else is there to support? What do you think it lacking?

  9. Jim Weingarten says:

    Eric,

    I worked at both VMware and NetApp, and was the technical liaison for the NetApp partner integration relationship with VMware. So, I imagine that we both have some valid perspectives.

    – VMware support for NFS has always lagged. behind block storage VMware refused to update their NFS implementation for years. They also refused to enable pNFS and NFS V4.X, despite the massive performance benefits those protocols offer.

    – VAAI forced all storage vendors to expose a common, controlled set of features, preventing differentiation between vendors. NetApp had a much larger set of capabilities than was made available to customers through VAAI. Additionally, VAAI and VVOLs were really a repackaging of native NFS functionality.

    – VSAN did not target any specific storage vendors, per se. It did offer a potentially compelling alternative to traditional enterprise storage. I agree that VSAN was an attempt to expand the value VMware could offer in its platform to maintain relevance.

    – NetApp’s feature set has been expanded. However, better, more cost-effective storage solutions have rendered ONTAP Edge irrelevant, in my opinion. While NetApp agonized over letting their IP out in the form of a VM, the market passed it by.

    Best,
    Jim

  10. wcurtispreston says:

    I have to say I agree with Jim here. VMware most definitely has its sites on the storage industry. Whether it’s because of the EMC relationship, or just because someone saw the handwriting on the wall, but it’s pretty clear to me that they are now going after storage $$$.

    I visited EMC the same week I visited NetApp and here’s what I wrote about that:
    https://storageswiss.com/2016/03/30/vsphere-tackles-the-hyperconverged-infrastructure/

    • ericmb says:

      evening Jim and WCP,

      I agree, there are massive changes going on. I just think all storage companies face the same challenges, not just a few. At the end of it I reckon it ll be a case of: the more things change, the more they stay the same.

      Also, I think all 3 of us agree that VMW are taking on all industries in an attempt to sit firmly in between and sement itself. With that in mind I think companies like Nutanix and their approach is interesting.

  11. Phil Gibson says:

    NetApp has an installed base to leverage and a revenue stream to support themselves. These are a couple of good steps as hardware makes a performance difference but the future is the hybrid and the cloud architecture. The acquisition of Steelstore is a great addition. Their best move now would be to separate their software from their hardware and to make a solid play as an SDS vendor. A radical move but services are a safer future than storage hardware.

  12. Jim Weingarten says:

    I think Phil has a point. In the end, NetApp will have to give up most of the revenue generated by its hardware sales. That means NetApp becomes a software company; something it has claimed to be anyway. Unfortunately, that also means the loss of the hardware margins that have been very high, historically. NetApp may survive, but it will be a much smaller company, if it is not acquired in the meantime.

Comments are closed.

Enter your email address to follow this blog and receive notifications of new posts by email.

Join 22,236 other followers

Blog Stats
  • 1,553,940 views
%d bloggers like this: