Yes, it’s true. EMC really IS #1 for VMware.

I REALLY try to refrain from chest beating – I don’t personally think it’s needed, and furthermore people don’t generally respond well to it.  

Sometimes it’s hard – as I’m fortunate to have access to reams of data and other information that isn’t widely known – so when I hear a strange claim by a competitor, it’s hard to sit back.   Usually, I do.   I just sit back quietly as others make claims that I disagree with…  

…But, today, I’m going to make an exception.  

We’re going to take a detour into clearing the air before returning to your regularly scheduled Virtual Geek nerd technical dialog, and into the realm of market share, analysts, studies, and the EMC/VMware dynamic.  WARNING: Stop reading now if you’re not interested.

If you are interested, and want to see me support it with piles o’ data, read on…

It drives me bananas when I hear all sorts of people say “look we’re more integrated with VMware than EMC!”.  This may seem weird to you, but I have to deal with it all the time.  Sigh.  Analysts, press, and listening to other vendors in VMworld sessions, heck, when vendors hold their public analyst sessions.

Well, I’m no Bruce Banner, but I AM about to Hulk out on you!

So – here’s the scoop.  They do it for several reasons:

For a while (think the 2003-2008 period) it was TRUE – at that time I think EMC tended to be a little confused on the topic of VMware.

1. Key attributes of the relationship (that are STILL TRUE!!!) - “independent” (which VMware still operates 100% independently) and “keep at arms length” (in both directions) were confused by EMC and VMware in those early years nto “don’t partner aggressively as critical technology partners should”.  That less than ideal dynamic happened in both directions.
2. In those days the degree of engineering integration was existent, but was light.
3. Our support for the VMware and VMUG community (which is in VMware’s case, “community” is very important) was non-existent – EMC was effectively a “no show”.  At the time, EMC’s interaction with their customers tended to be “high-touch”  (direct customer engagement), vs. the low touch-model of VMware (using VMUGs and channels).
4. For a while our “Go-To-Market” was way, way off when it came to VMware.  There was no “why EMC for VMware” deck.  There we no answers from the field to the question of “who do I call for help”?  I still cringe thinking back to the first VMworld for me in my current role @ EMC (was VMworld Europe 2008 in Cannes).  I had just gotten put in the role a few weeks prior – so I was walking into a situation that was unchangable (everything was locked and loaded).  I knew it was going to be bad, but I had no idea HOW bad.   The EMC booth had no demos, no hands on, and the stations were named by EMC product (which most VMware admins wouldn’t know), not VMware use case.  We had one session, which was fluff, and wasn’t even in the session calendar.

Ugh – it was a mess.

There’s another reason why other folks do throw those easy catchphrases out… because it’s juicy.   It’s an easy dart to throw (people love suggesting that there is infighting between a close family – and people love to speculate).   People also do it because it’s easy to poke at a EMC as the leader in our space where we are the 800lb gorilla.   A nice, friendly, innovating gorilla, but a gorilla nevertheless.

They also do it because if they are having success (and many are for many reasons in areas where we need to improve – EMC is far from perfect – leaving room for competitors to have successes), and can link it in some way with VMware – even with little or NO proof, it drags you up in the market.   We all live with a “pump and dump” market – which gives me a bit of “frisson”.

For the last 2.5 years of my life, I have been pouring all my willpower and energy into the EMC and VMware dynamic – and ALL of those 4 items above are NO LONGER CORRECT IN ANY WAY.   I’m setting the record straight.

Looking back at the list of 4 reasons:

1. We are not confused anymore – we are CRYSTAL clear on all fronts.  While VMware remains independent and open (partnering in the field openly, and sharing APIs with everyone), EMC has and will continue to out invest in every area (field, marketing, near term product development, longer term - significant parts of our $2B long term R&D) – winning the hearts and minds of VMware customers, VMware partners, and VMware themselves on our own merits.   
2. The degree of engineering integration is now overwhelming.   I challenge any one company to show as much as we do across so many fronts (storage, security, backup, security).  These range from “independent projects” (think vCenter plugins), “do a better job than anyone on keep up with the jones’ projects” (think things like SRM, VAAI, or VADP – stuff the whole vendor community needs to eventually support), to “intimate engineering projects” (think PP/VE that like the Nexus 1000v is fundamentally a vmkernel module - and many, many other secret projects that exist in various states).
3. Our support for the VMware community is overwhelming – you can’t shake a stick without hitting a vSpecialist, we have a universal VMUG support (we will sponsor, provide any content/demos needed, with a commit to make it technical and not a vendor commercial), and in the blogosphere, people are now saying EMC might be contributing TOO much.
4. Our GTM is focused, and at VMworld, VMUGs, vForums – we are pervasive and omnipresent.  We try as hard as we can to keep it focused on technology - though some inevitable marketing fluff seeps between the cracks (which I try to stomp out).    At the VMworlds this year – we had 2.5x the booth space of our competitors, more kit supporting the HoL, and about 3x more sessions.    When it comes to joint customer interaction, we represent a big selling partner of VMware’s doing several multimillion ELAs in the last couple of weeks alone.   We have hundreds of public reference customers – more than 30 of which of which joined us at VMworld on stage and in the booth sharing their stories.  From a field resource standpoint, there are almost 200 vSpecialists (with more every day) focused on helping customers get value out of EMC and VMware, and THOUSANDS of EMCers who have various levels of VMware certification.

THAT’S NOT CHEATING, IT’S JUST EXECUTION.

What do I mean?   Look at the sessions at VMworld as an example where it comes down to simple execution.   Here’s the inside scoop on how we did it.  When someone is a Diamond sponsor, you get a given number of sessions (for example, this year, it was 3).   Yet EMC was in 21 sessions at VMworld Americas (some solo, some with VMware).  That was because of number and quality of the session submissions we made that the community voted for.  As soon as VMware opened up for session submissions, I leveraged the whole company to submit proposals in time, and en masse.   Success/failure rides on a knife edge, and the little execution things make the difference.

The EMC competitors see this and are starting to realize they will not be able to out-invest and out-focus us in this area – so are starting to say “well, we support Hyper-V and Xen!”   EMC supports everyone of course – we know customers want choice (including the choice to have integrated infrastructure like Vblock).   I’ve done many posts on what we do around Hyper-V (example, example) and Xen (example), and we will continue to invest proportionately to customer demand, and partner furiously with Microsoft and Citrix.

I’m also not saying that others don’t integrate well with VMware.   I’m not saying we don’t have competition, heck very strong competition.  I’m even not saying that some are ahead of us in some ways – usually small, and usually transient – but that’s the nature of the game.

But… A lack of clarity around the overall success of EMC in the market?  That’s STOPPING.

We had these ALL over both VMworld events this year, and you’ll see them around the world in many places – airports, cabs, magazines – you name it.   And, we’re just getting started.  You’ll not see us let up.

We’ve also simplified things for our field – with a simple “Top 10 Reasons Why EMC for VMware” battle card (click on it to download).

And hey – people around the world speak different languages, so here are the PDF versions:

Chinese, Japanese, Korean, Italian, French, Portuguese, German, Spanish, Russian.

Yes, of course, anytime you simplify things down to a sentence, there are LOADS of details and nits with each of these (which I’m sure our competitors will start to disassemble) – but every one of them is defendable.

So – what am I saying? 

1. THANK YOU customers for choosing EMC.   You can count on ongoing and overwhelming EMC efforts on all fronts to support your VMware journey.
2. THANK YOU EMCers and EMC Partners.   I thank you for all your continued efforts around everything we do.
3. THANK YOU VMware. I thank you for your ongoing world-changing innovations and continued openness to partnering corp-to-corp and field-to-field.   We partner with you with humility and openness, and thank you for doing the same – together we help our customers.
4. THANK YOU competitors.   You have my absolute respect.   Together we support VMware and our cusotmers, and our competition is good for customers and keeps innovation fast and furious.   But – I won’t stand by and let people misrepresent the state of the market, or denigrate EMC without support for their argument.

Unlike the things I tend to see out there – I’ll put validation behind what we say.  Want proof?  Read on.

OK – first, let’s talk about sources for the “EMC #1 Storage for Virtualization”…

It’s VERY hard to get this metric consistently but there have been a variety of studies over the years.   Goldman Sachs (here), IDC (here), and others.   One of the more recent ones is the ESG Server Virtualization Study (data collected July 2010).   BTW – AFTER READING THIS, GO BACK TO EACH, LOOK FOR THE PATTERNS ACROSS IDC, GOLDMAN, ESG STUDIES.

 

For EMC folks and EMC partner who would like to use these slides – download them here (and all the rules on how you can use them here).

Now, a couple key observations – this is a relatively small number of customers, Americas only.   To extrapolate out from that to claim “market share” would be INCORRECT.   ESG is firm about the statistical accuracy of the results – the conclusions are correct in the sample, just don’t extrapolate incorrectly.   “Primary vendor” was defined for survey respondents as the storage vendor supporting the greatest percentage of virtual server capacity.   And yes, the totals may not add up to 100% due to rounding.

I’m sure those are the points that EMC competitors will point out.  

Well, we don’t use the ESG report to track our progress internally.  

For that, we do a VERY large scale study across many countries and geos, and we are doing it quarterly.   This is from a firm that doesn’t have the “3rd party analyst” branding of ESG or IDC or Gartner, but they (Management Insight) are used extensively by VMware, Cisco, EMC, and VCE for studies to help us with internal decision-making (products, roadmaps, priorities, pricing, support – a million things) by helping us get real customer feedback at scale.

Because I’m SURE our competitors are discounting ALL the studies, I thought I’d give a public peep into this study – and again, I CHALLENGE PEOPLE WHO CHALLENGE OUR CLAIM TO SHOW A SIMILAR SCALE STUDY.

Here’s the scoop:

• Timing: Data collected July-September 2010
• Approach: Web-based quantitative survey of IT professionals who are involved in purchasing decisions for storage for virtualization
• Sample Size: 632 US, 609 Europe (210 UK, 197 France, 202 Germany) – BIG & BROAD.
• Audiences:
  • IT Decision-makers and IT Implementer
  • Enterprise and commercial organizations – BROAD SEGMENTS
  • Using virtualization and storage technologies
  • Familiar with storage for virtualization
  • Aware of at least one brand of storage for virtualization
• Topics:
  • Inventory and Attach Rate
  • Awareness, Consideration, Preference (ACP), and Brand Ratings
  • Last Purchase Case Studies
  • Switching Dynamics
  • Software ACP and Topical Questions
  • Profile/Classification

A little more info on the “who was asked” (note that people of course don’t use CIFS with VMware, but it can still be a “deployed technology”)

What were the results?   There were a LOT.   A lot of good insight on what people like, and what they want us to do better on (the primary purpose of this ongoing study).   But – I’ll pull out one summary from the data analysis.

BTW – compare that chart with the first ESG one – the one that has the not specific to VMware, but rather "server virtualization”.   Do you see a pattern?   The scoop is that in general, there is a a strong bias towards the same vendor who sells them the server.    It’s no surprise – but that’s why IBM and HP are where they are – it’s certainly not on the basis of VMware integration leadership in the storage platforms.  But, aside from the server vendors, EMC and NetApp preference pops up – but EMC overwhelmingly.  

In the internal study, when the question is asked more specifically as “Storage for VMware” – the EMC percentage pops up from 33% up to 47%.   Where have you seen that before?   Look up at the ESG charts again (and the Goldman/IDC ones).

So – does this correlate with marketshare gains/losses?  Not exactly.

While VMware is growing like mad, it’s still just a small part of the storage world.   Certainly the fastest growing use case, or host attach, but marketshare gains or losses are drowned out by other macro trends like:

• the ongoing growth of unstructured NAS for both EMC and NetApp outside the traditional NAS use cases (applies to startups too)
• Channel dynamics – very much about margin, services, and other non-technical questions.
• Shift towards multiprotocol, multi-tier platforms (more “in the box” becoming the ante)
• Healthy, but not massive iSCSI growth (still not “exploding”)
• the mass deployment of disk dedupe targets
• small overall growth, but major marketshare shifts at the high-end (towards EMC)
• Increasing use of cloud object storage models (EMC and non-EMC like Amazon) at the expense of traditional NAS
• Service providers are big, but we’re still in early days of cloud compute – enterprise/commercial business

Also it’s worth pointing out: there’s little factual support for any “architecture” being right/wrong for VMware.  In fact, it may be be a statement of of the blazingly obvious, but SMB customers tend to consider either the SMB offerings from the server vendors (or things like iomega), mid-sized customers or enterprises virtualizing phase I apps (things that aren’t mission critical) tend to go for mid-range class arrays (like EMC Unified), and those virtualizing mission-critical apps tend to put them on platform architectures they’ve always tended to select for their most mission critical apps (like EMC VMAX).  These are mass observations, there are exceptions of course.  

IMO, anyone saying “this architecture is BUILT for VMware” is trying to sell you something.  If you’re a customer, they are selling you a box, if you are an analyst, they are trying to convince you to pump or buy their stock – note EMC is guilty of this too – not saying we are guiltless.

In that context, while EMC being #1 for VMware is awesome (which EMC is), it’s an important, not the only determining factor in EMC’s (or any other storage vendors) growth.  

Well – how does this correlate with marketshare gains and losses (the IDC data everyone in storage land lives and dies by)?   IMO - it has a loose, but not a direct correlation.   It’s interesting to see how people have mashed up this data against the broader marketshare gain/losses generally tracked by IDC most broadly.   You can see one such “mashup” here: http://wikibon.org/wiki/v/Projecting_VMware_Storage_Market_Shares:_EMC_Continues_to_Lead

So why does everyone seem to insist on making these outlandish statements?  Saying “we’re winning the cloud wars!” “we win more than X when it’s VMware” is a sound bite analysts understand.

Ok – so what about the WHY DO I CARE AS A CUSTOMER?   What about data to support the “60 integration points”?

You care because integration is good.  It makes things simpler, more efficient.   “Integration” was in the middle of the pack in the Management Insight study details over why people pick one storage platform over another.   BTW – in the mass study of customers, integration was the first desired feature after the classic requirements (“it’s gotta perform, it’s gotta be available”)

So – where are we on integration?   This is not the exhaustive list, but let’s take a quick look.

The idea of a list of these things (developed based on customer need, not to “make a list”) is kinda silly if you ask me (as it’s invariably used for stupid comparison “check box lists”).   We try to track internally for our own uses.   When the marketing folks asked us for “how many integration points can we definitively prove” we went off and counted.   It’s MORE than 60 – and I won’t list them all, but I’ll put down some of the highlights (not claiming that any of these are UNIQUE, but that the SET is the most comprehensive).

• EMC Unified vCenter Plug-in (Provision, configure, and extend VMFS and NFS datastores, Monitor VMFS/NFS datastores, Mass-replicate individual VMs on NFS, datastore level on VMFS, integrated with VDI brokers)
• VM-aware Unisphere (Insight into vCenter from Unisphere)
• Virtual Storage Integrator (Monitor block storage datastores, Provision, configure, and extend block storage datastores on EMC Symmetrix, Mass-replicate individual VMs, Discover and identify EMC arrays, see multipathing relationships, automates many SRM setup steps on Symmetrix)
• EMC Unified/EMC CLARiiON auto-registration driven from vSphere 4.1
• EMC Unified/EMC CLARiiON SATP in vSphere 4.1 has customized queue and tresspass code
• EMC Unified/EMC Symmetrix T10 Thin Provisioning Stun behavior with vSphere 4.1
• Replication Manager integrates with not only all sorts of apps, but with vCenter and ESX for VMware-integrated point in time local and remote replicas on all EMC replication products – with one simple tool.
• Data Protection Advisor vCenter Plug-in
• EMC Control Center vCenter integration
• Iomega v.Clone
• Ionix Unified Infrastructure Manager
• Ionix Data Center Insight integrates with vCenter, and has a vCenter plugin
• VPLEX vMotion over distance
• EMC Unified/EMC CLARiiON VAAI Support today (VMAX coming in Q4)
• PowerPath/VE
• RSA Envision Integrates with vCenter, ESX, View
• RSA Key Manager SDK for vCenter Integration
• RSA SecureID Integrates with ESX, View, vSphere Management Assistant
• RSA Archer integrates with vCenter, ESX (and heck everything under the planet) and future Intel Westmere TPM TXT/vSphere
• Avamar – integrates with vCenter, VADP
• Broad SRM support (Celerra Replicator, Mirrorview, SRDF, Recoverpoint) with very mature SRAs.
• EVERY EMC platform has a failback-plugin that extends parts of failback beyond what SRM does today
• CLARiiON MirrorView Insight
• Recoverpoint integrates with vCenter directly, showing the relationships with replicated VMs directly.
• EMC products publicly available as Virtual Appliances:
  • Atmos Virtual Edition (prod and play)
  • Avamar Virtual Edition (prod and play)
  • Celerra FileMover Appliance (prod and play)
  • Celerra VSA (play)
  • NetWorker FastStart (prod and play)
  • RSA Data Loss Prevention Network (prod and play)
    • Network Controller
    • Network ICAP Server
    • Network Interceptor
  • Solutions Enabler (prod and play)

Two quick notes:

1) I didn’t (and don’t) list things that aren’t real integrations, but are real examples of “affinity” (where we do something that helps in the VMware use cases but isn’t technically integrated) – for example, the effect of FAST Cache for View, FAST’s ability to deliver DRS for storage in conjunction with SIOC today, the ability to support NAS and block use cases, or VPLEX and vMotion/VM HA stretched clusters.

2) I’m on the warpath to SHORTEN the list (while it also continues to get longer as more stuff comes).   As an example, there loads of these our customers aren’t getting enough leverage out of.  For example, we’ve just gone down from 3 storage vCenter plugins for storage down to 2 (VSI and the EMC Unified Plugin), and then shortly after to 1.  We shouldn’t have 3, we should have one.   In doing so, we’ve worked such that all EMC vCenter plugins will integrate into that common model, making it easier on our customers.   Yes – some parties are ahead with “one plugin”, but conversely none have the breadth of customer requirements we serve. And… For each area were we are not first, I’ll point out an example where we are – firsts tend to be transient.

…. OK.   Phew.   Perhaps this will help no one, but helped me psychologically to put it to proverbial paper.  

The next time someone says “we’re more integrated than EMC”, or “VMware prefers us to EMC”, or “we’re gaining marketshare behind VMware” – remember saying something is as easy as breathing, and there are a LOT of “less than brilliant” mouth-breathers.    Ask them to PROVE it, and remember that proof is not one or two anecdotal customer stories, but DATA.

So, back to your regularly scheduled technical Virtual Geek programming…   Before leaving, please comment – is this useful (i.e. you’re hearing people making the kind of claims I’m talking about), or am I off my rocker?

Comments

[BrandonJRiley]

Good write-up. I hear this out there quite a bit. I think the study you show from ESG is indicative of the current state of IT research these days. I can show you an Aberdeen report that shows almost the exact polar opposite of what the ESG one shows.

Of course the Aberdeen one doesn't specify demographics, or even number of respondents. I do think it illustrates why we all should be as skeptical of these "research" firms as we are vendor marketing. Is there really a difference?

[Chad Sakac]

Thanks Brandon - is the Aberdeen report publishable, or someplace where you can link to it?

The ESG data on it's own would be one thing, but when lumped with all the others, it gets a little more weight, and ultimately the non "3rd party" study is actually the one I think has the most weight as it's the broadest sample size.

Still - they are all worthless in any given customers' decision making. They should pick the solutions from the vendors and integrators that are the best (in their own opinion as the customer) suited to solving their particular challenges, and from the vendor/partners that have served them well in the past (experience counts for a LOT).

Chad

[Oldmanaround]

As of 10/4, RSA enVision now has full support for VMware vShield Zones/App/Endpoint, as well.

[Jason Boche]

Great read Chad. I am amazed by you passion to crank out these chapters on you blog as busy as you must be.

[BrandonJRiley]

Chad:

I sent it over via e-mail. It's one of those documents that's locked away behind their "information firewall". When you get a chance to read it, I would enjoy hearing your thoughts.

[Andrew]

Completely appropriate and a lot of helpful data IMHO....if anything, I'd like to see this kind of thing from any storage vendor who claims to be the best "VMware anything" option.

There's no need for it to bash the competition....just cite all the strengths, positive reports, integration points, etc. -- let each storage vendor put their absolute best foot forward publicly (on a blog where they can control the comments no less too).

[Don Woodward]

Chad:

Can you write or have you write somthing about using Microsoft Clustering (MSCS) with the VBLOCK? It appears that MSCS only supports fiber channel storage - since the VBLOCK uses FCoE and Fiber Channel connectivity does this mean MSCS will not work on a VBLOCK?

[Don Woodward]

Can you write or have you written somthing about using Microsoft Clustering (MSCS) with the VBLOCK? It appears that MSCS only supports fiber channel storage - since the VBLOCK uses FCoE and Fiber Channel connectivity does this mean MSCS will not work on a VBLOCK?

[Vaughn Stewart]

Hmmm...

I'd like to request EMC make public the ESG server virtualization study from july 2010. I can;t seem to be able to purchase this report from ESG directly which raises more than a few questions.

http://www.enterprisestrategygroup.com/?s=server+virtualization+july+2010&submit=Go

How about standing behind the content in the report by making it available for review, so we can all believe that said claims "are true"

Cheers!