Blogging with Integrity?

Are there many things better than taking a vacation with family and friends? For me, one of the best parts of a vacation is the build up before the vacation starts. Knowing you’re work is wrapped up, the anticipation of the trip and time spent with the kids. Heck, I even get a little excited just telling friends that I’m about to start a vacation.

I spent last week on vacation with my immediate and extended family and it was great. The day before we left I even shared this news with a friend of mine, Chad Sakac of EMC.

In fact, I'd like to share with you the content of our email conversation...

Chad - Sun, 28 Mar 11:40 – I’m in Ashville, NC tonight, if you’re interested in dinner, and in the vicinity…

Vaughn – Sun, 28 Mar 12:28 – Ironically we are packing today for a vacation in the Appalachians and will pass thru Ashville tomorrow. Unfortunately I can't pop out. What are you doing in Ashville? Pleasure trip I hope.

Chad – Sun Mar 28 12:37 - Enjoy "hiking the Appalachian trail" :-) . (I’m) Giving them the "big picture" view from EMC along with VMware. Bummer I'm missing you, but enjoy your well-deserved time off.

(click on the image to view in full size)

As many know, I consider Chad Sakac a friend, even though we are fierce competitors. We talk or email with regular frequency, work on a number of joint initiatives, interact with the same customers, partners, vExperts, etc… Heck, look at the tone of our email conversation if you question our friendship.

Why Open a Post Discussing One’s Vacation Plans?

In recent weeks many have noticed a change in a number of posts by some at EMC when it comes to comparing the level of storage integration available within VMware vCenter server with arrays and technologies between NetApp and EMC.

Such criticism is fair; we’re competitors; however, I’m very disappointed that my friend Chad would level such accusations with full knowledge that I am on vacation and thus unable to respond to his claims.

Seriously, who does that?

If you spent time reading his recent posts it appears that he is working feverishly with a goal at eroding my credibility. I’ll save you from having to reference his posts by sharing the two claims, which he has levied.

1. I am presenting EMC technology in a negative manner
2. I am misinforming the public in regards to both EMC and NetApp technology

It is clear that both of these claims stem from the following chart, which compares vCenter integration between storage arrays from EMC & NetApp.

(click on the image to view in full size)

A Bit of Background

During a number of customer visits in August of 2009 a new theme had emerged in the marketplace. A number of these customers had stated their EMC accounts teams were making the claim that EMC, as the majority shareholder of VMware, had an inside track to engineering resources and as such provided “unmatched completeness of vCenter integration.”

This claim prompted customers to request more details around the level of vCenter integration provided by both NetApp and EMC.

As a member of the NetApp virtualization business unit I work directly with VMware engineering and this claim didn’t align with programs like VAAI where VMware, NetApp, Dell, and EMC jointly work on defining storage integration standards. The idea that VMware provides integration points for individual storage partners is outrageous.

I have invested a significant amount of time reading EMC product manuals and presentations, viewed numerous EMC video demonstrations and attended multiple EMC webcasts. As such I have obtained a significant amount of knowledge as to what EMC can and can’t do and as such I disagree with their claims around leading the market in terms of vCenter integration. This is when the comparison chart began.

On To The Heated Debate

In Chad Sackac’s recent comments on my credibility he claims that he, and for that matter EMC, takes the higher ground when consulting with customers in that they plainly talk about what EMC can do versus what the competition can’t. I can tell you from first hand knowledge, some of which includes Chad himself, this claim is far from the truth.

To be clear, every IT vendor investigates the capabilities of their competitors and positions against them. This is standard practice of any competitive industry.

Being driven to respond to this banter generated within field engagements around the world, I presented this chart in a two part blog series titled, “EMC: The Storage Most Integrated with VMware?” Here's the links to part 1 and part2.

In part 1 of this series I included the sources of this data contained within the chart. I noticed EMC has excluded any references to this post and sources in their claims that the represented data is incorrect. This omission seems odd and lacking the transparency that I am advocating.

 

Digging a bit deeper into the claim that I am misinforming the public…

In part 3 of my blog series, “VMware Admins are Storage Admins - vStorage Integration”, I clearly state my position around the purpose of the vCenter integration chart:

"My intent here is to be very accurate and fair in representing the non-NetApp technology. My intent is to help clarify the pubic understanding. If I've overlooked a detail, please let me know and I will revise this content."

I believe this statement articulates a position to compare technology with input from the community.

As one might expect the post drew some criticism around the content of the chart. All of which understandably came from employees of EMC. Following through with my desire to provide correct information I reached out to these individuals in both email and in the comments section in order to obtain further clarification around their concerns in order to accurately depict the data contained within the chart.

It may surprise some to note that the chart was updated within 24 hours of the original post based on these communications.

Fact Finding

I’d truly suggest that anyone interested in the complaints from EMC, read the comments sections of part 3 in the “VMware Admins are Storage Admins - vStorage Integration” series. I have posted Chad’s audit of the chart as it appears in the comments section. He reviews each of the 15 lines on the vCenter integration chart (each line is numbered in his comments) and in reproducing this audit I’d like to complete a public review of the criticism. In this manner we can track the correctness of the content of the chart and/or his comments. I’d ask you to do the same.

Again, what follows is Chad Sakac’s review of the vCenter Integration Chart (as it appears verbatim in comments section) followed by my comments and score keeping.

1. Auto-Provision Datastores – (The chart is) correct.
   
   Score
   Chart 1 – Chad 0
   
   
2. Dynamically mask LUNs - incorrect. There is a plugin for Celerra that does that.
   
   I cannot find any content on an EMC plug-in that auto-provisions LUN based datastores for Celerra. As such I will not score. If EMC can provide a link I will adjust the chart. - Verdict incomplete.
   
   Score
   Chart 1 – Chad 0
   
   
3. Dynamically grow/shrink datastores – (The chart is) correct.
   
   Score
   Chart 2 – Chad 0
   
4. Dedupe VMs - putting an exclamation mark in there is fine if it makes you feel good :-) We both save capacity, in different ways. EMC's F-RDE is a combo file-level dedupe and sub-file compress. In the VM datastore use case (which we're talking about here), file-level dedupe does nothing, sub-file compress generally nets a 40-50% capacity savings (in general purpose NAS use cases, file-level dedupe often saves more than block-level or compress). It has the advantages of having NO impact on filesystem size, features, or behavior, and being unaffected by local and remote replication (ergo, there is no "pinning" for elements of a filesystem that are being referenced by a snapshot). Can the same be said of the NetApp approach? Not saying better/worse - just pointing out that a NetApp constructed table, from a NetApp constructed world-view won't note the pros/cons on both sides - just like an EMC-constructed one would it?
   
   I would like to point out for completeness, that EMC claims to support dedupe for production VMware Virtual Machines, when as stated before it is not dedupe, it is compression. Apparently, transparency includes the redefining of terms in order to position to be included in a technical evaluation. We dedupe any data, including VMs, with any storage protocol; EMC compresses VMs on NFS with Celerra. < Chad, this is a HUGE difference.
   
   However, in trying to be fair we did award EMC a yellow exclamation point while many would argue that it deserves a big red flag, we will let the customers decide.
   
   Score
   Chart 3 - Chad 0
   
   
5. Report dedupe savings - same as above. I personally would say "report space savings", and then it's incorrect.
   
   While, Chad is asking for modification to the criteria maybe I should update the chart to reflect “Report Space Savings.” Doing so again blurs the lines between what is dedupe and what is compression. I’m open to customer opinions on this topic. It would seem to me that you would want to know the method by which you are gaining storage efficiency. NetApp has many storage efficiency technologies, dedupe is only one of them and the purpose was to underscore the effort we have placed in integrating savings achieved through the use of dedupe to the VMware administrator.
   
   Score
   Chart 4 – Chad 0
   
   
6. Auto-configured iSCSI settings – (The chart is) correct.
   
   Score
   Chart 5 – Chad 0
   
   
7. Auto-configure NFS settings – incorrect, it's auto in the Celerra NFS plugin.
   
   I believe I might know more about EMC plug-in technology than Chad, maybe one of his vSpecialists can update him. To be very clear here, the Celerra NFS plug-in does NOT auto-configure NFS settings, it allows the admin to select which settings they would like to have set. There is no assurance that the correct settings are being made, just the ability to make settings. Again, we listed this capability as a yellow exclamation point. In contract NetApp provides the Auto-Configuration of our best practices which are cooperatively developed with and certified by VMware engineering. We do so to ensure our systems are connected following best practices. Again, the purpose of the chart was to underscore our engineering work which simples the provisioning process. Providing an interface consisting of multiple choices is not automation. For now I’ll score this one to NetApp and if the feedback supports changing it to a green dot I will do so.
   
   Score
   Chart 6 – Chad 0
   
   
8. High Performance Multipathing - incorrect, see above.
   
   I don’t understand Chad here. We all know one can use the native multipathing in ESX, but PP/VE is available for a reason right? It must provide some value at a price point of a few hundred dollars per CPU socket.
   
   I’ll quote implicitly from the EMC website, "PowerPath/VE enables you to automate optimal server, storage, and path utilization in a dynamic virtual environment. This eliminates the need to manually load-balance hundreds or thousands of virtual machines and I/O-intensive applications in hyper-consolidated environments."
   
   I believe a green checkmark with PP/VE is a correct mark for high performance multipathing I/O as PP/VE provides “optimal server, storage, and path utilization”
   
   Score
   Chart 7 – Chad 0
   
   
9. Physical to virtual storage management – incorrect. iSCSI and FC are supported across the board, the Celerra also has it for NFS.
   
   I’m not sure this claim is correct, as I cannot find any content on the EMC plug-in that provides iSCSI and FC on all platforms. As such I will not score. If EMC can provide a link I will adjust the chart.
   
   Score
   Chart 8 – Chad 0
   
   
10. Array based VM cloning - this is incorrect, but on your side. This can only be done on NFS datastores, not VMFS datastores. We both can "cheat" (taking a snapshot/clone) of the LUN, mount and copy out a VM, but that is not a VM-level snapshot. The reason we're able to do it on NFS is that both Celerra (as of DART 5.6.47) and ONTAP (I think as of 7.2.3?) can do file-level snapshots (which in the VMware use case manifest as "VM-level" snapshots). Is this statement of mine correct?
    
    Ok, there’s a lot going on in Chad’s response here, so allow me to break it down…
    
    - Can EMC and NetApp create VM level clones on NFS datastores? - yes
    - Can EMC and NetApp create datastore clones? - yes*
    
    *NetApp for any protocol
    
    *EMC with iSCSI and possibly NFS; however I cannot verify the NFS claim with any online content. If a link is provided which validates the claim I will correct.
    
    - Can EMC and NetApp create VM level clones from datastore clones? - yes
    
    *NetApp for any protocol
    
    *EMC with iSCSI and possibly NFS; however I cannot verify the NFS claim with any online content. If a link is provided which validates the claim I will correct.
    
    In summary, Chad is partially correct. Under array based VM cloning the NetApp rank should be more explicit as to how we handle VMFS. I’ll blog on this after my vacation ends. Score one for EMC.
    
    Score
    Chart8 – Chad 1
    
    
11. Array based datastore cloning - incorrect. Celerra can do it for NFS and iSCSI.
    
    See above, the data is not available to be publicly confirmed as being a capability delivered from within vCenter. -- Verdict incomplete
    
    Score
    Chart 8 – Chad 1
    
    
12. IO/offload for VM clones - incorrect, on your side. See 10.
    
    Here’s the details around cloning for a number of deployments...
    
    - Single VM clone on NFS – array based
    - Single VM clone on VMFS – ESX/ESXi clone
    - Mass VM clone on NFS – array based VM clone followed by array based datastore clone
    - Mass VM clone on VMFS – ESX/ESXi clone in one datastore followed by array based datastore clone
    
    So partial credit to Chad, the VMFS operations do require a form of host based cloning, however, when cloning in mass, as with VDI, the ESX/ESXi cloning only occurs on a single datastore and as such the majority of I/O is offloaded.
    
    Chad is partially correct here, but also misleading with his implication as you can see this criteria is vague.
    
    I’ll give this one to him and revise the chart.
    
    Score
    Chart 8 – Chad 2
    
    
13. SRM support – correct
    
    Score
    Chart 9 - Chad 2
    
    
14. SRM failback - incorrect. it is supported on all four - in the latest SRDF SRA, Recoverpoint 3.3, Mirroview Insight, and the Celerra SRA.
    
    This claim is correct. The data in the chart represents what was available with version 1.0 of the EMC SRM failback plug-in. EMC has since enhanced their SRM failback offerings. My error, I will update the chart
    
    Score
    Chart 9 –Chad 3
    
    
15. Not sure if I understand the last one - can you explain to me what it means?
    
    Certanly, NetApp’s SnapManager for Virtual Infrastructure allows its replicated backup sets to be used by SRM for DR test and DR failover. One copy of data, stored offsite which serves a dual purpose. To my knowledge EMC can only provide this capability with Recovery Point and not with other backup technologies like Avamar.
    
    Score
    Chart 9 – Chad 3

Maybe we could move this comparison to a third party, say prominent bloggers in the virtualization and storage space. Chad, would EMC sign up to co-sponsor such an event?

Keeping Score and Maintaining Perspective
 

Alas it looks as If my chart requires some minor updates, and without further adieu, here’s the chart updated to include the claims by Chad Sakac which could be validate. I would like to formally request others to substantiate the chart and any outstanding claims (complete with publicly available links).

(click on the image to view at full size)

Frankly, I don’t find the changes very compelling. I guess the biggest changes are around adding SRM failback to 2 cells (out of 105) in the chart (see the Symmetrix DMX and VMax columns).

This update is exactly why it is so important for customers to see the difference in the level of vCenter integrations from NetApp and EMC. NetApp offers a single platform that supports small, medium and large workloads. That single platform is a long standing Unified Architecture which has been shipping for years. NetApp integrations with VMware are across our entire Storage Array portfolio. You need a chart, a wiki page and 10 different manuals to figure out if this EMC platform combined with another has a plug-in to deliver features X or Y. It’s pretty simple over here at NetApp, we release integration through innovating technology in collaboration with VMware, it is available to all of our customers, on all arrays, with every protocol, and not a just select few.

Availability of vCenter Integration with Vblocks

I don’t expect EMC to be happy with what I’ve shared as it doesn’t present a compelling story around the integration with their three discrete storage array architectures. I also don’t expect them to endorse this revised chart; however, I do believe it is accurate and it would be nice to see EMC focus on what they can do versus what would like you to believe that we can’t.

Even when one implements all of EMC’s VMware integration technologies a VI admin still cannot deploy a datastore or reduce the total storage footprint on a Symmetrix or Clariion array, and these are the arrays positioned to host enterprise deployments with EMC Vblocks 1 & 2.

(click on the image to view in full size)

Now EMC will tell you they can through using a VMware feature, like linked clones but is that an innovation in an EMC storage array, nope its an innovation available to any VMware qualified storage platform. Is it accurate to claim rights to an innovation in your feature when it is available on any storage platform?

Based on the amount of vCenter integration on the Celerra and the number of concerns one has to ask, is Celerra the new flagship for VMware on EMC deployments?

The answer appears to be ‘YES’ if one wants some features on par with what is available from NetApp.

Chad, help me understand. It appears one could obtain most, if not all, of the integration capabilities with Symmetrix and Clariion with arrays from HP, HDS, Dell, etc… Is my observation correct?

Closing the Loop

I’ve taken every means to act in as transparent a manner as possible. I’d love to say that I’m acting as an advocate for the consumer, but that wouldn’t be honest because I am having to respond to the aggressive sales tactics of EMC masked under the claimed veil of honesty and integrity. As a former Systems administrator and IT Manager I have an appreciation for the ability to receive accurate and concise data. I believe I have presented the VMware vCenter integration capabilities of NetApp and EMC in such a manner.

I’d like to ask you for your thoughts. By sharing, discussing and explaining the technical differences between what a manufacture says and the reality of what the technology delivers are these efforts helping to educate the VMware community or am I misleading as some claim?

This is a great discussion and I welcome all feedback

As for Chad – he might want to re-read the Blog with Integrity charter, looks like a few have slipped from his memory.

Comments

[Chrisdbirdwell]

All of the information that you are looking for is publicly available Vaughn. I am surprised you had trouble finding it. I am on vacation as well, but will be happy to provide links next week when I get to the office (If none of my fellow vSpecialists provide it first).

[Vaughn Stewart]

@Chris - That would be great. Enjoy your time off and ping me when youre back to work. Im reachable at [email protected].

This message was sent by my thumbs... fault them for the brevity and poor spelling

[Chad Sakac]

@Vaughn - apologize that I wrote that post in anger.

To be frank, I didn't get the "look NetApp presented at the MN VMUG!" email until after you and I had our email dialog - and got me infuriated. I can see how that timing was hard on you while on vacation, but that wasn't on my mind at the time - again, my apologies for that.

This whole episode has reminded me that to blog while angry is not a wise thing, just like emailing while angry.

I stick to my guns on one thing (along with still being angry). It's also not wise to directly throw competitors under the bus, and I think it reflects badly on you and on NetApp.

My point was the table was incorrect (you may think it's not material, I think it's material - in the end, the customer chooses), and incomplete (ditto), and the process of keeping it up to date is always going to be incomplete.

Likewise, any similar competitive exercise where EMC tries to keep "more current" than NetApp on NetApp invariably results (to the rest of the world and customers) with EMC looking bad, as it's impossible to know a competitor better than you know yourself. (do you REALLY think "I believe I might know more about EMC plug-in technology than Chad, maybe one of his vSpecialists can update him." - seriously?

BTW - On point 15, thanks for being more clear, so let me be more clear. EMC Replication Manager does what SMVI does (local replica that can be replicated to a remote target) across EMC replication technologies (in a way that I believe is similar to the way SMVI leverage Snapshots and Snapvault/SnapMirror operations at the array level). If that's how a customer is best served to do backup, great - we can both position that as a solution.

In some cases customers don't view (for many reasons) array based snapshots (even those replicated offsite) as a substitute (sometimes they DO view them as an augment for fast recovery under some circumstances) for enterprise backup (aka NetBackup, TSM, Commvault, NetBackup/Avamar to a target) as a backup solution. BTW - this is why EMC and NetApp fought so furiously for Data Domain. One answer is not always the right answer.

I've come to the sad conclusion that the only choice I have is to do a one, by one, line by line video recording of all the things claimed about EMC that are incorrect, supporting my original post comments, which I will do, of course - after EMC World. I have my hands a little full. Sigh.

The other element you selected to cut from my comment was that of course, being a NetApp constructed slide (designed from your world view, and designed to make you look good, of course), it neglects a series of vCenter plugins EMC has that NetApp does not. For example, Avamar integrates vCenter for backup reporting and management, and integrates with CBT and VADP. Is there a NetApp analogue? What about DPA's integration with vCenter, providing multitenant backup visibility? Is there a NetApp analogue to that? What about SCM's integration with vCenter, is there an analogue to that?

My point in saying that isn't that one is better/worse - it's that I think these marketing oriented things are stunts, and have limited value.

The key thing to me at least is that this is a recurring conversation with you and I - one that we've had many times over (starting with that first time in VMworld Europe when you crashed the EMC session - which you're always welcome to do). My viewpoint is that NetApp is a fine company, and has fine technology. They are do themselves a disservice when they directly start with negative comparison approaches. Literally, when training new vSpecialists on how to compete, my comment is "it's a very good thing that you can count on NetApp to, unprompted, throw EMC under the bus".

On to the other point (unrelated, but whatever) - likewise as the original points, Vaughn, you're incorrect. Both Vblock 1 and Vblock have NS-G8 (analagous to vSeries) as recommended elements on the BoM. All vCenter plugins available are supported on the Vblock. Like any product, Vblocks themselves (which should be viewed as a product, not a set of products) have an integrated lifecycle and product roadmap. Customers can expect ongoing updates to Vblocks, just like any product.

I hope that the whole episode didn't taint your vacation. It certainly tainted my week.

[Vaughn Stewart]

@Chad - thanks for the apologie.

As I've said in the past, EMC makes great arrays, but I believe NetApp technology is more in line with the requirements of a virtual data center or cloud deployment.

The chart demonstrates my point.

As to expanding the chart, I could but does it really help EMC? In all seriousness a VI admin can't even deploy a datastore in vCenter for a Symmetrix or Clariion array. Sure, one could buy a Celerra with their vblock 1 or 2, but vblocks 1 & 2 are designed to serve VMs on Symm and Clariion LUNs and not Celerra.

Any claims that suggest Celerra fronting a Symm or a Clariion should publicly state that VMware on Celerra is EMC's preferred storage architecture.

Are you willing to make such a statement?

Of course not.

I'd like to ask you to encourage correct sharing of information and drop these personal attacks. I don't believe they position EMC storage technology in a favorable light.

Please don't take out your frustrations for the lack of EMC & VMware integration on me, it would be better targeted at EMC engineering.

We are going to compete for the foreseeable future, let's try to make it a little more professional as we go forward shall we (even if one of us has a less than stellar offering).

[Jason Linden]

Vaughan & Chad,

All I can say is wow. I read Chad's post last week and said, "Rut Ro, this is going to get nasty." Then there was near silence on twittersphere. "Hmmm, NetApp must have really been launching hand grenades and running away." Of course I now understand why NetApp was quite.

I recently, within the past 6 months, joined the vendor side after spending the past 10 years in enterprises. Most recently I was at a company as VI 3 was being released and led the charge to virtualize the entire organization.... On NetApp storage. It was fabulous, besides the installation engineer not properly cabling filers and caused an outage when we had a hardware failure - almost lost my job (but we won't go there). In all honestly, I really enjoyed the NetApp platform and bet our business on it.

Fast forward to Nov 2009, I leave to join AdvizeX Technologies as the Consulting Practice Manager for Virtualization. How great, I am an independent mind and no ties to hardware, right? Wrong. AdvizeX is a top 10 EMC solution provider (and a top 10 HP solution provider, but we won't talk about their storage platform here). It could be the continual infusion of Kool-aide, but I have come to like the EMC product line up.

Why do I say all of this? I understand the war you guys are in. I know the FUD being tossed by both sides. Frankly, I am just glad to be watching a VMware related war and it not be about Cisco UCS. Won't go there. Anyway. I am more disturbed that you both claim to be friends. Friends don't let friends get stuck in the mud in blogosphere or twittersphere.

Chad mentioned in his post he has attempted to resolve the issue personally without success. I certainly hope this is true. I find it quite disturbing how far off the deep end this conversation has gone. After watching this for a bit last night, I wake up this morning to TweetDeck full of more banter and RT to just to make sure everyone knows your still fighting. You guys aren't the marketing guys, so stop treating it like a marketing conversation just pointing out what makes your product look better.

Personally, I think NetApp & EMC (or you guys personally) should jointly fund Tolly (picked them just to stir up the UCS focus) to do your comparison for you because neither of you will be objective at this point. One would hope the "friendship" would drive objectivity, but that has surely driven less objectivity. Once you guys build the matrix together, then come back with a joint post about features and stop the battle.

Better yet, open the June 11th VMmug with a hug and jointly speak to an agreed upon strengths/weakness chart. This is what our customers want! They are sick and tired of the continual vendor banter! Step up and your lumps were deserved. If you guys can't do this, then rent a UFC ring for the VMmug and resolve it that way...

Blog with Respect & Blog with Integrity. They are not mutually exclusive...

[Vaughn Stewart]

Re: [NetApp - The Virtual Storage Guy] Jason Linden submitted a comment to Blogging with Integrity?

@ Jason Well said. Customers deserve credible information regarding the abilities and capabilities of the technology they are considering to purchase and I fully endorse your suggestion to have multiple vendors or 3rd parties flesh out content like the vCenter integration chart.

My wife asked me why it was so important to have this conversation and in my reply I used the following analogy. Say your looking to purchase a new car and of the models at the top of your list one has a turbocharger and the other a supercharger. You know both enhance the vehicles ability to accelerate, so is that enough information for you to select the type of booster which meets your driving requirements? I suspect most would say no and at some point ask a sales representative for more information. Each sales representatives would share that their technology is a forced induction system that compresses the airflow in an engine. Now is this enough information to make a decision? For some maybe, for other its not enough and at some point someone has to go compare the two different technologies designed to provide similar results.

Id like to think of comparisons like in these posts as information sharing. Will the charts age and become inaccurate - yes. Will we have minor errors from time to time yes. Yet even with these flaws, we can start discussion which allows the data to be refined and ultimately in the hands of all.

Thanks again, I hope the negative personal tones which have infiltrated many blogs in recent weeks will cease.

Oh and for the record, my car has a supercharger.

[the storage anarchist]

Vaughn -

I'm restisting getting knee-deep into this argument, but allow me a couple of comments on the Symmetrix columns. With a little more info from you, I may back to offer some adjustments:

First, I think you have been perhaps a bit liberal in the NetApp check-marks in some of the categories. Maybe I misunderstand your point, but some of the first few rows seem to be specific to using VMware on top of NFS and unrelated to how VMware works with block storage.

That said, Symmetrix V-Max provides features commonly referenced as "auto-provisioning," while both V-Max and DMX provide "dynamical LUN masking" and even "grow/shrink" devices (thick and thin). Being an unabashed block-head, I'm just not sure we're talking apples/apples on these.

And on that note, giving Symmetrix a Red X for not being able to "Auto-configure NSF settings" seems a bit unfair, being as Symmetrix is a block device. As you later note several "NFS-only" features for NetApp, it would probably be more fair to say "N/A" for the NFS-specific stuff for both V-Max and DMX (and CLARiiON, for that matter).

You've also got me confused on the array-based "VM Cloning" and "datastore" cloning - depending on how you're using "datastore," I could argue that all of EMC's block storage platforms provide array-based datastore clones.

And the last one, "failover to backup disk array" has me confused as well - why do you say "RecoverPoint only"? Unless I miss your point, SRDF provides this functionality as well.

Like I said - if you can help this block-head understand what you're trying to say, I think I can back up some more changes that will Get (some of) The Red Out!!!

(Oh, and it goes without saying. If I were creating the comparison list, I might have a few more/different rows in the table. But like I said, I'm trying not to engage the battle, just clarify some of the points being made).

[Nick Howell]

brb grabbing some popcorn. :)

[Kent]

Vaughn - perhaps you could alter the chart to demonstrate the feature support across protocol stacks? Right now many of the features or points you make are only available in certain protocol stacks, and all protocol stacks can not necessarily be implemented in all environments. Some environments are Fibre Channel for a reason, so a chart that ignores the differing feature set across protocols might lead to incorrect analogies or inferences.

So as an existing NetApp customer utilizing the FAS3020 in FC environment, I might gather that I could benefit from things like dynamic grow/shrink datastores, which I obviously can't in my configuration.

[the storage anarchist]

How odd - you've now removed the "NFS Only" tags from the NetApp pages after I and a NetApp customer pointed out the same disparities...and changed apparently without any explanation.

"Blogging with Integrity?" - methinks not.

[the storage anarchist]

My apologies - I opened the wrong (old) version.

Still, Kent and I are waiting for your response as to which of the other features are NFS only and thus inappropraite to be compared with the block-only devices.

[the storage anarchist]

Crickets...

Must be something I said.

[Vaughn Stewart]

@Barry (aka the Storage Anarchist)

Thank you for sharing your thoughts and time, I’m truly flattered. The premise of this chart is to compare the level of storage integration made available to a VMware admin through VMware’s vCenter server. I believe some of the folks with storage backgrounds sometimes overlook this criterion.

Regarding the top few rows of the chart, specifically the LUN content, we can do exactly what is stated in the chart. From within vCenter a VI Admin can provision a net new datastore and our plug-in will complete this process including the LUN provisioning, masking, setting attributes like thick/thin, selecting paths, setting the multipathing policy, enabling ALUA, and connecting the LUN to all of the hosts where it gets formatted with VMFS.

The plug-in even handles the differences between VI3 and vSphere hosts, even when these two platforms operate in a mixed cluster.

On you point about NFS with the Symm… I think you're right. I plan to revise.

On datastore clones… again the chart depicts what a VI Admin can accomplish within vCenter. I'm not challenging that a storage admin can clone a LUN, but there are a number of additional steps required after that one in order to turn a LUN clone into a usable datastore. Would you agree?

On backups… NetApp has a very unique backup solution for VMware. We provide snapshot-based backups of VMs via SnapManager for Virtual Infrastructure. These backups reside on the production array and exist as the first line of data recovery.

As a best practice, backup data should to be copied onto another set of media and preferably stored offsite. When one replicates the NetApp snapshots to another storage array this remote offsite copy is not just a backup set but it can also be used for DR failover with VMware’s Site Recovery Manager.

Dual purpose of one set of data, is pretty cool (and saves $$$).

This capability is rather unique and not found in many disk to disk backup solutions (like Avamar or Data Domain); however, RecoveryPoint does provide this capability. As you know, EMC offers many backup solutions as hey provide various benefits and functions. RP is a CDP and replicates every block of data on every system, as such; it may not be an ideal for all of one’s VMs due to the bandwidth requirements.

With NetApp we deduplicate the production footprint, replicate the deduped backup data, and host offsite backups that are deduped and usable with DR. With your background and understanding, I hope you would consider this model impressive.

OK, I'm getting into 'sales mode' a bit here.

I'm flattered that you'd share your thoughts and comments. If you'd like more info or even a demo, where you drive, just let me know and I'll make it happen. Ping me at [email protected] and we'll go from there.

Thanks again.
V

[Vaughn Stewart]

@Kent

Thanks for the comment. I like the idea. I am planning on revising the chart and offering EMC to have input on it. Let me see if I can make your suggestion work.

Cheers!

[www.facebook.com/profile.php?id=658313066]

Vaughn, thanks for taking the time after getting back from your family vacation to construct such a balanced, salient point by point technical rebuttal.
Thanks to you both for challenging each other to cut through the marketing and continuing to help the customers create the best solutions.

-- Netapp and EMC customer