« It's Happening Again | Main | Charting the VMware Revolution »

February 24, 2009


Alex McDonald

Difficult formatting this thing, but given the top of the range CX4-960 and the same functionality as the FAS6040A you've quoted, without disks, the result is;

Base Unit CX4-960C (*ESTIMATED PRICING*) CX4-960 - SPE 124,000
Software VP4-960 (*ESTIMATED PRICING*) CX4-960 - Virtual Provisioning 73,000
Software NAVAYZ4-960 (*ESTIMATED PRICING*) CX4-960 - Navisphere Analyzer 80,400
Software NAV4-960 (*ESTIMATED PRICING*) CX4-960 - Navisphere Manager 115,000
Software NQM4-960 (*ESTIMATED PRICING*) CX4-960 - Navisphere QoS Manager 83,500
Software MV-CEM4-960 (*ESTIMATED PRICING*) CX4-960 - MirrorView Cluster Enabler (CE) 13,800
Software MVA4-960 (*ESTIMATED PRICING*) CX4-960 - MirrorView Asynchronous 70,300
Software MVS4-960 (*ESTIMATED PRICING*) CX4-960 - MirrorView Synchronous 93,300
Software SV4-960 (*ESTIMATED PRICING*) CX4-960 - SnapView 46,400
Software SC4-960 (*ESTIMATED PRICING*) CX4-960 - SAN Copy 82,000

$781,700 total.

That's also without a ton of storage, and let's be kind and assume that the Celerra part of the NS-960 is free, and there's no cost for the CIFS or NFS or iSCSI software, and there's definitely no guarantee, and that the Celerra could actualy virtualise other vendor's storage.

Maybe I'm slow, but what exactly is your point?

Chuck Hollis

Alex, you must be standing way too close to this one.

It's simple -- the latest offer from NetApp is "plug in a v Series to third party storage and save a bunch of money".

I'm just asking the question -- exactly when do you start saving any money with this approach?

Nice job on the CX4-960 config, though ...

-- Chuck

Alex McDonald

I think *you're* standing way too close. Take your nose off the pricelist, and look at the issues customers face with sweating their existing SAN assets. Multiple SANs from a variety of vendors, looking all the world like pools of DAS. Then think of a solution. And it's not an EMC rip and replace strategy folks are looking for.

The V Series guranatee is part of reducing risk and cost. The other savings by using a V Series can't be matched by EMC.

Oh, NetApp customers got dedupe for free ages ago. On NAS and block level SAN FC and iSCSI too. That's the beauty of having a unified storage architecture instead of a unified storage marketecture.

Stephen Foskett

Ooh ooh ooh! We have pricing! So since neither of you said "you're way off on that price" can we assume you're NOT way off on that price?

My hobbyhorse this month is cracking the veil of mystery storage pricing. Just so's ya know...

Mike Shea

The Buzz from VMworld Day 1:

...and at the end of day 1, the VMware hands-on lab is running on our NetApp FAS3070s. EMC’s 15 boxes are still not working. Well, maybe tomorrow…

Sorry man, that was a jab...

Thinking out loud - perhaps if we put a V-Series in front of them you'd have better luck...

(Hey - just having a little fun...)

Chuck Hollis

Hi Stephen

List pricing and configurators are out there already, what's really interesting to find out is how different vendors do street pricing, optimize for different configs, bundle, etc.

Gartner has made a good living for years coaching customers on how all this works. Of course, for the low-end NAS stuff, there's always Froogle :-)

If you crack the code, please let us know.

-- Chuck

Chuck Hollis

Alex, it's a simple question, isn't it?

I'm not looking for rhetoric, just a plain answer.

Proposition from NetApp: spend X and save Y on storage.

What's the breakeven point? Any wild guess? Did anyone check that out before writing the breathless press release?

(answer: probably not)

-- Chuck

Chuck Hollis

Oh, BTW, Alex, we think we've figured something else out as well, maybe you can comment?

The original NetApp "guarantee" (??) was 50% against a RAID 1 config using RAID 6.

Now the newer "guarantee" (??) appears to be something like 35% against legacy storage.

We wondered where the 15% went.

Answer: it's the WAFL tax -- duh.

-- Chuck

Alex McDonald

The breakeven point for V Series can be 0TB or less. If you were running an ISP with VMware customer images, then being able to pack more systems onto your existing SAN (or perhaps you might even contemplate a mix of SAN and NFS, or decommisioning small older systems) can enable your growth without requiring yet another SAN or expansion of existing, possibly at capacity, systems. Other savings; power requirements over larger systems, reduced cooling and floorspace, etc make it an attractive proposition.

Your choice of the V6040A was at the very high end of NetApp's range of V Series. Most customers will be looking to the V3000 series, which provide exacty the same features and functions as the V6040A but at a lower price point.

The V Series program covers a wide range of existing EMC, HP, HDS and IBM equipment, all with differing RAID types. Because we are integrating multiple SANs from a single V Series that may be other than RAID-1, there will be more disks for data than the equivalent RAID-1 configuration.

Hence the %age savings are 35%. How you can characterise a significant 1/3 *saving* as a "tax" is beyond me.

Chuck Hollis

Alex, thanks for trying, but it just comes across as more nice marketing words. And they're not even good ones.

So, let's try this again ...

Implementing a v Series gateway (or any similar product) has a defined cost -- both for the product, and in the effort to get it all working.

Let's call that investment "X".

The product is now directly positioned as "use less storage" with all of the ancillary benefits that you point out, and several more as well. Storage and the infrastructure has a certain cost per unit for the customer.

Let's call that number "Y".

All customers will want the savings from "Y" to be significantly larger than "X", right? Otherwise, why bother?

Let's just make up some numbers to see how this works.

Imagine a customer said "you know, I've figured out that -- all in -- I spend about $1 per GB per year to keep this rotating rust spinning".

NetApp says "well, my esteemed marketing people tell me I can save 35% of your capacity" (once the WAFL tax is figured in). But it's gonna cost you $50k per year (all in) to do this.

At 50TB of capacity savings, you'd break even, right?

50TB of savings implies that you started with 142TB, forgetting for just a moment all the other scary implications of running that much capacity through a filer head.

But customers aren't interested in breakeven propositions, are they?

So the numbers have to get much bigger, don't they?

And as the storage numbers get bigger and bigger, the cost of the required NetApp equipment get larger, the effort required increases, and the chance for it all turning into a sorry mess increase dramatically as well.

For 100TB of capacity savings, the customer would have to start with close to 300TB of initial capacity, requiring a bigger, beefier set of v-Series devices (according to your specs, that is). So the cost and the effort go up dramatically.

Now, my numbers aren't real ones, are they?

We did the best we could to run the numbers, and found that -- under no circumstances -- would the customer ever get close to breaking even. Let alone coming out ahead.

Call us biased or whatever. You need to prove us wrong, because this latest stunt makes your company look really, really bad.

If you guys really cared about solving customer problems, you'd have an answer to this question.

You don't.

Alex, my good friend, didn't your parents ever tell you that actions speak louder than words?

Because it's pretty clear this wasn't about solving customer's problems. It can't be. Even on the back of a napkin, the numbers don't come close to working. We all spotted it in about 30 seconds.

That's because the motivation is quite different: it's just yet another NetApp marketing stunt in a very long series marketing stunts.

That's OK, that's your choice, isn't it?

If you'd like more ideas on future marketing stunts (expensive rap videos, annoying banner ads everywhere, "guarantees" that mean nothing, manipulating the SPC, etc.) I have several suggestions for you to consider.

Personally, I think that's a great way for NetApp to spend their time and money.

-- Chuck

Alex McDonald

YMMV. Several thousand V Series customers would disagree with you.

One further request. Please leave out the personal references when replying to me. You know nothing about my background, my childhood, my parents, or anything else of a personal nature.

I post on your blog and uncomfortable facts get up your nose. That's it. Deal with it like an adult. It would be much appreciated.

Lee Razo


I notice you tend to talk about NetApp and use phrases like "I just don't get it" quite often.

I think the way most of our V-Series customers end up saving money is when the support contract on their EMC gear expires and they are finally able to migrate to a full NetApp solution.

Not only do they save tons of money over years by finally getting to use NetApp in their data center, but with less time troubleshooting and fine-tuning messes like the Celerra Franken-NAS and Franken-VRLs and the like, they also get to spend more time with their families, exercise more regularly and eat better - making them happier, healthier and more attractive to the opposite sex.


Lee Razo

@Steve Foskett:

EMC will almost certainly come out _far_ cheaper than the equivalent NetApp solution. In fact, if you want to get rock-bottom prices from EMC, just tell them you're talking to NetApp - or better yet, that you have some NetApp gear in your data center.

I also noticed that Alex was kind enough to leave off the the services pricing from the EMC price-list. EMC has been known to leave that off the quote too. But I dont know of any customer who would actually want to assemble that beast without any help.

Geoff Mitchell

Assuming that NetApps has several thousand customers, can you post from this sample at least a couple of instances with the numbers where the value proposition is demonstrated? I think Chuck has posed a valid question and I would like to see more of those uncomfortable facts.

I get the concept of storage virtualization and see value in having this in several topologies. Although I've seen a couple of installs where this is an enabling technology, I've yet to see an example where the V-series is purely a cost saver on the spinning stuff.


Alex McDonald


On our website you'll find many; here are a few.

http://media.netapp.com/documents/entara.pdf pair of HP XP12000s

http://media.netapp.com/documents/swisscom.pdf HDS SAN

http://www.netapp.com/us/library/customer-stories/jcv.html EMC CLARiiON, Symmetrix and NetApp FAS systems

Lee Dilworth

Update from VMworld. Hi I am responsible for building out the kit used for the SRM lab. I just wanted to take this chance to say a HUGE thanks to both EMC and Netapp for providing storage and hands on assistance in terms of people (Alex / Luke a big cheer for you guys couldn't have done it without you).

From an availability and performance point of view both the Clariions and the 3070s have been working very hard to support a mulitutde of labs. Each pair of arrays is handling well in excess of 50 replicated luns EACH (both sync and async) as well as many other non replicated devices acting as the VMFS volumes for the server VMs.

From the lab team point of view we could not have been more impressed with our storage platforms, surviving the trip down here in the back of a truck is one thing and then performing this well day in day out is something else.

We haven't lost a single disk, never dropped a single replication session and each array must have by now performed hundreds of snapshot tests.

thanks again to all involved from EMC and Netapp.

Chuck Hollis

Thanks, Lee -- glad we could help!

From all indications, VMworld was a fabulous show with great impact ... keep up the good work!

-- Chuck

Shannon McMahon

Stunt or not, the analysts like it. One example: http://esgblogs.typepad.com/

Chuck Hollis

You'd have to know Steve pretty well to understand why he tends to like that sort of thing :-)

Then again, I could point to Beth Pariseu's post calling the "guarantee" hollow ...

-- Chuck

James Sillence

Hey Chuck

This blogging warfare always makes really amusing reading and I love how easy it is to get the NetApp fan boys fired up.

What's more interesing however is that thankfully, customers do look a little deeper. I have used the "Space Saving Guarantee" from NetApp very successfully in campaigns and have influenced the decision against them. It all goes to credibility. Once a customer recognises that 43% of the saving can be atrributed to RAID choice - the game is up.

This latest announcement will be no different. I too don't understand when a customer will start to save real dollars. What I do see is a way to increase complexity, management overhead and invalidate support agreements. As far as I see it, it's just another tool for me to use when speaking to customers who have "real world" problems to solve.


Chad Sakac

Lee, you are a class act my friend.

Mike Shea - do you mind communicating internally at NetApp that the message you posted (which looks like an internal message - at least to me) is completely, and utterly factually incorrect?

If you don't want to - that's fine, but claims like this that are easily corrected don't help your cause, they undermine your credibility.

Luke (NetApp) and Alex Tanner (EMC) worked hard to make VMware suceed here. Anyone who has supported an event knows that there are ALWAYS last minute crises, and both NetApp and EMC had them (I know, I am here).

The fact that they both pulled crazy hours to resolve and both performed flawlessly at the event to support VMware is an example worthy of respect, not dishonorable positioning and falsehoods.

BTW - one thing that Lee didn't say, as he is a good man - there are 5 EMC arrays (not 15, though we have another 4 arrays in our booth, not including a bunch of ix2 and ix4s) and the 2 NetApp arrays. The CLARiiONs support not only replicated LUNs but also non-replicated LUNs in the SRM HoL and also served as a backup target for much of the HoL, and substantial other parts of the show.

I will be posting a bunch of photos (including the HoL) and videos from the event, as well as session content. In the end, we had 5 EMC sessions, 2 additional joint VMware/EMC sessions (Sharepoint on VMware best practices + VDI and DR).

Of those 7 sessions - 2 ranked in the top ten (of 300) by attendees (Sharepoint and one EMC one - TA11), and were asked to be repeated.

There are demonstrations and more to post, but need to sleep - it's 2am Cannes time, and I need to wake up at 4 to catch my flight :-)

Alex McDonald


The latest 35% offer would appear to fly in the face of your statement that 43% of the saving in the original offer can be attributed to comparing RAID-DP (NetApp's RAID-6 equivalent) against RAID-1. There's no requirement to compare with any RAID configuration; RAID-5 is fine, or even any other vendors RAID-6 from what I can see.

I'm interested that you believe that putting a V Series in front of, say, an EMC CLARiiON, would invalidate support agreements. Can you explain under what circumstances that might be the case, and whose agreement might it invalidate?

Thanks in advance.

Chuck Hollis

Alex --

There seems to be considerable confusing as to the exact terms, limitations, exclusions, etc. of the various "guarantees".

As an example, I have been sent at least 6 different documents (all with handsome NetApp logos) detailing slightly different variations of the infamous "guarantee".

Would you mind sharing links to the authoritative descriptions in use today? I think everyone would appreciate it!

-- Chuck

Alex McDonald

Chcuk, you can find them in the right hand box "How You Benefit" at http://www.netapp.com/us/solutions/infrastructure/virtualization/guarantee.html . The last revision is dated February 2009; and the Best Practice guides can be found there too.

Chuck Hollis

Thanks, Alex -- basically, the same structure, exclusions and limitations as the previous version(s) we've seen.


-- Chuck

James Sillence


That's true Alex, RAID choice isn't a factor in the latest guarantee - I wonder if we will see the original guarantee reissued. By my maths, if the vFiler can achieve 35% with similarly configured RAID, then without using the RAID slight of hand in the original guarantee, you should be getting close to 78% right? But then I guess you really would have to lengthen your list of caveats and exclusions.

As for support - if hosts are connected to a device from vendor X that sits in front of a storage array from vendor Y - who assumes end-to-end responsibility when the customer has problems?


Alex McDonald


Invalidating support agreements sounded much more serious than who assumes end-to-end responsibility when the customer has problems!

It's an unusual data centre that has all brand X; we've all, over the years, accommodated to the fact that most customers don't have, and don't want to have, all their eggs in one basket. Like every vendor, we'll make every effort to help the customer, much as if he'd connected server from vendor X to vendor Y's switch to a NetApp FAS array.


Chuck, IMHO your rethoric is way to easy.

You say: "[...] That's because the motivation is quite different: it's just yet another NetApp marketing stunt in a very long series marketing stunts."

Shannon says: "Stunt or not, the analysts like it." -linking to Steve Duplessie, who is in turn downplayed by Chuck: "You'd have to know Steve pretty well to understand why he tends to like that sort of thing :-)"

So Chuck, what about Gartner? They're supposed to be discredited the same way when they publish a research paper named "Capacity Savings and Storage Utilization Guarantees: What's Included, and Are They Worth Considering?" (ID Number: G00162398, http://www.gartner.co.uk/DisplayDocument?ref=g_search&id=791312) that says:

- While some customers are wary of anything that looks like "marketing hype," campaigns such as these are not inherently bad. As long as there is contractual recourse for not meeting the guarantee and users enter into the programs thoughtfully and with a full understanding of the program and its remedies, these programs can be quite beneficial.
- If the company is going to guarantee capacity savings, it makes sense that it would require users to implement the features that will allow it to do so, and NetApp gets credit for being so transparent about the requirements.
- I think very positive specially compared to EMC, HDS and HP. Because non of them are offering similar programs.
- When researching this report, Gartner contacted several disk array storage vendors to inquire whether they had a capacity savings or storage utilization guarantee program or had any plans to institute such a program. EMC and HP indicated that they did not.

So I'd say to any customer should take Gartner's recommendation to heart when they say: "During the research phase of your next storage purchasing cycle, ask each vendor if it offers a capacity savings or utilization guarantee."

If the answer is "NO", anyone should wonder why...
At least, that's what I would do.

Chuck Hollis

Hi Sjon

I think vendors -- all vendors -- should be held accountable to deliver stated results. There's absolutely no argument whatsoever on that point.

However, constructing an artificial "guarantee" that (a) has dubious technical merit (e.g. use RAID 6 as an alternative for what once ran on RAID 10), (b) has numerous exclusions for databases, email, etc. etc. -- well, that just struck many of us in the industry as a marketing stunt, rather than an attempt to deliver honest value to the customer.

And I wasn't the only one. ...


I think Sjon, Shannon, Steve, Gartner, a whole bunch of analysts and us at NetApp do not agree with you and whoever it was you weren't the "only one" with...

So I guess whether something is a marketing "stunt" is - among beauty - in the eye of the beholder....

No authority on this subject but the market itself, right?

Chuck Hollis

You're right. It is in the eye of the beholder.

And when I ask serious IT users about the "guarantee", they roll their eyes and start chuckling. They think it's about as funny as the time NetApp did SPC tests on EMC kit.

Because -- of all the people you mentioned -- they're the only ones who really matter, right? And -- if anything -- it looks like NetApp is keeping them amused.

-- Chuck


So you think I am not a serious IT user, Chuck?

I always respected EMC for it's vision and enjoyed reading your blog but I think I just got personally insulted.

If this isn't the way you treat your "unserious" IT customers/prospects I think you may want to think twice before speaking....

That said - you may want to down play their "stunt" all you want, at the end of the day the market will decide. Like it decided on the viability of NAS before....

Chuck Hollis

OK, Sjon, take a deep breath ... I didn't mean for you to take it personally.

Everyone is entitled to their opinions: me, you, everyone.

I shared what I had heard directly from other IT users I had met. I don't understand why you'd be so offended by other people's opinions, but that's another story altogether.

Thanks for writing and sharing ...

-- Chuck

Gene Piatigorski


I have reached the same conlusion when an EMC Sales Rep. was trying to convince me to get an Invista. Has EMC performed the same ROI analysis (which is missing a whole slew of costs - monitoring/reporting tools, products support, technology skills/resources, etc.) on their own storage virtualization products?

My current perspective is that storage virtualization over-promisses, under-delivers, and does not provide an ROI that would support an acquisition, from *any* vendor. At least I have not yet found one...

Maybe EMC is willing to give away its virtualization products to make the ROI attractive? After all (in my mind) it is the design/operation of the SAN storage array that creates the necessity of having a vitualization layer to ease the management processes. Should it not be *baked* into the Tier 1, Enterprise-level solution like a Symmetrix? From a customer perspective, that was a rhetorical question...


Chuck Hollis

Hi Gene -- thanks for writing.

Invista should have been positioned as enabling hot movement of workloads between arrays. If there's not a value prop for that, it might have been mis-positioned by someone.

And you're right -- as with any storage virtualization approach, there's a slew of other costs (direct and indirect) that should be considered but often is not.

As far as "baking in" -- there are some array-based features for hot data mobility between arrays (DMX's Open Replicator combined with PowerPath Migration Enabler come to mind as a popular one), but the capabilities are specific to the array, and not available at the generic fabric layer -- to each their own.

Thanks again for writing!

Gene Piatigorski


Regardless of the positioning of the products, does EMC determine the pricing of their solutions based on an expected ROI by the customer? In my experience I would say no. Invista is only one example. I can rant about many others...

In my own analysis (with the EMC Sales Rep., we did include the use of Invista for data migrations as a potential benefit. It still did not cut it...

Are you saying that the DMX Open Replicator is available to customers free of SW license costs? This would be what I would call 'baked in'. If a customer has to pay for it, it is not 'baked in' in my mind... The other issue with your example is that PPME sits on a host and requires care and feeding more than the benfits it potentially provides in your example. I would hardly call this an array 'baked in' feature from that perspective.

Thanks for taking on the hard questions from a customer. It shows that your blog is not ONLY about marketing hype and assasination of your competitors' marketing ploys. It shows that you are willing to see the wood log in your own eye while observing the dust spec in your compteitors' eyes...


Chuck Hollis

Gene -- all very good observations.

At some point, I'm hoping you and I can sit down for a very rich conversation ...

In the meantime, thanks for sharing.

-- Chuck

Michael Hay

Chuck, I think that in light of the Data Domain purchase discussions on this post some come back alive again. I've asked some pretty simple questions as a purchase of this company seems to put some additional holes in the argument as now to get additional capacity savings on Data Domain you have to shell out more cash... Thoughts?


The comments to this entry are closed.

Chuck Hollis

  • Chuck Hollis
    SVP, Oracle Converged Infrastructure Systems

    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!