Chuck Hollis of EMC on NetApp marketing techniques.
Click here or read below.
If you promise a customer that your solution is more cost effective, it shouldn’t end up being more expensive. If you say your solution is simpler, it should not be more complex than what it replaces. If you say its high availability, it should be. If you claim superior performance, that is what it should deliver. And so on.
A few weeks ago, I took the gloves off on a particularly egregious example of Bad Marketing, where NetApp had pushed the limits of benchmarketing too far; to the point where it could probably result in the exact opposite of what was promised.
Now, to be clear, I have no problem with head-to-head competition. I welcome it. It makes all of us in the vendor world better, and it ultimately benefits customers. And I love the rough-and-tumble, he-said-she-said world of IT marketing. It’s fun.
Today, I’m going to go a bit further on NetApp on some of their other marketing claims. Not because I don’t welcome an aggressive competitor (I do, believe me), but I cringe when I find people actually believe what they’re saying, and get hurt in the process.
The premise is that – because NetApp has one primary operating system (ONTAP), and one generic hardware architecture (dual controller storage), that customers can serve all their storage needs with a single, compatible family, saving considerable operational expense.
Now, if I’m a smaller shop, and my requirements aren’t that demanding, I’d give NetApp the benefit of the doubt. I certainly have met smaller customers who could conceivably run their entire shop on NetApp if they limit what they do.
But let’s look at the facts. There are tens of thousands of customers who’d never consider putting a critical database on an emulated SAN device like NetApp – it’d kill their performance. What happens when a small customer needs that capability, and has to put something different in to do that job?
As an example, all NetApp customers are now being “strongly encouraged” into a very disruptive upgrade to ONTAP 7G. Usually, there’s new hardware involved, and it involves substantial downtime, unless you use something clever like Rainfinity.
Are there any meaningful performance tuning tools in ONTAP? Not really, that would add complexity. Want to do VTL? That’s a different NTAP storage platform. Want to do encryption? Decru isn’t integrated. Want to do CDP? That’s Topio, also unintegrated.
First, they point to thin provisioning as using less storage. Actually, thin provisioning helps solve certain poor storage management practices that are common in IT (see here for more info) but comes with its own – ahem – complexity.
Second, they point to a low-overhead snap capability that doesn’t actually make the second copy until it’s written to. Nice trick, but also available from other vendors, including EMC. And, of course, there are pros and cons with different use cases when this would make sense, and when it wouldn’t. It’s not a panacea, and it’s not unique.
Nothing wrong with having competitive features compared to other vendors. But to claim that you use less storage than the other guys basd on these claims simply isn’t true, and isn’t supported by the evidence.
First, as many people have noticed (and I’ve pointed out) the only reasonable way to address a performance problem on a NetApp filer is to add more unused storage. When we get called in to solve a performance problem on NetApp, we usually notice that the customer has taken this step, and is using much more storage than would otherwise be needed, simply to present more spindles to ONTAP and WAFL.
Second, EMC has built a nice practice of going into large NetApp shops, running some simple analysis, and showing customer by either deleting or archiving the gorp that’s sitting on their filers, they can use 40-70% less storage.
Put differently, because NetApp doesn’t offer the tools or the practices to do this to their customers, we find that unmanaged NetApp shops use far more storage for file storage than is usually required.
After reviewing some of the NetApp marketing materials, you’d think that Oracle wouldn’t be able to run on anything else. I run the alliances group at EMC, so I have a bit of an insider perspective on this.
NetApp spends big bucks every year sponsoring Oracle Open World. Well over a million dollars a year, I believe. Oracle really appreciates this money for their event, and reciprocates by saying very nice things about NetApp.
NetApp also did a great transaction a while ago with Oracle’s Austin Data Center where they run their outsourcing business. I’m not privy to the exact details, but there seems to be an exchange-in-kind where Oracle got a great deal on hardware, and NetApp got a great marketing reference. The result is a slick video where Oracle people running Austin talk about how great NetApp is.
So, what does Oracle run their production systems on? The ones that keep their company running? Their email? The majority of their development environment? The gigantic Oracle Single Instance at the very heart of their company?
So, back to the theme. I’d agree that NetApp is Oracle’s preferred marketing partner in the storage world these days, but there’s a far stretch between that financial arrangement and the one NetApp is claiming.
More specifically, do companies trust NetApp with their most mission-critical applications? Do they have the robust architectures, backed up by expensive service and support capabilities? Do they have experience in some of the most demanding IT environments in the world?
Usually it’s test and dev environments, or maybe a large file farm, or something else that isn’t deemed mission critical. And, trust me, they do a very good job selling to places where they’re good, and staying out of environments that they don’t do well in.
But – please – don’t torture that statement into positioning yourself as an enterprise storage company. You folks haven’t made the investments in products, testing, service and support to back up that claim.
Now, to be fair, I’m picking on NetApp pretty hard here. They’re big boys, they can take it. Heck, they dish out plenty of dirt to EMC, so they must expect a bit of a rebuke regarding their business practices. And the practice of Bad Marketing is hardly limited to NetApp.
I sit on several marketing review committees, and we try – as hard as we can – to make sure that if we claim something in our marketing material, that we can back it up, and – more importantly – we’re not leading people into the exact opposite outcome. We’re not anywhere close to perfect in doing this, but we consider it very important to make a consistent effort in this regard, and to try to get better.