Attack Of The (Vblock) Clones

I feel a bit guilty invoking Apple and iPads, but -- let's face it -- the company pushed out a wildly successful new computing device, and now everyone else in the industry is scrambling to react.

In our own little corner of the IT infrastructure market, the same sort of thing has been happening.  

Converged infrastructure -- in the form of the Vblock from VCE -- has become enormously popular with enterprise IT, partners and  service providers alike.

Not surprisingly, in a competitive industry like ours, something that's been so successful must be responded to, and Vblocks are no exception.

During the last year, we've seen several attempts to clone the considerable successes of the Vblock.  

None of these attempts appear to have gotten much traction so far.  At least, they're never mentioned by customers and prospects.  There's also been a predictable amount of FUD from vendors who don't have the ability to make an equivalent offer and would like to dissuade you from considering one.  

That apparently hasn't been successful either.

So, what's going on here?

And what do competitors need to do to make an equivalent or superior offer?

 
The Allure Of Integrated Infrastructure

The idea is simple.  Take the very best from VMware, Cisco and EMC.  Combine it into an integrated platform based entirely around private cloud concepts: built differently, operated differently, consumed differently.  Sell and support it as a whole.  Enlist a wide ecosystem of partners. Point it at leading IT organizations who want to evolve their IT infrastructure in a big hurry.  

Once a Vblock was fully understood, the industry split along predictable lines.  

On one side were the traditionalists who preferred to stay with the build-your-own approach to IT infrastructure.  On the other side were more progressive IT types who realized that times had changed, and it was time to move to the next thing.   The acrimonious debate continues to rage.

The great part?  A Vblock is nothing more than a new consumption option for existing advanced technologies.  If you want to buy individual piece parts and assemble / integrate / support them yourself, that hasn't changed -- have at it.  

Or you can consider a Vblock if you'd like -- the choice is yours.

Indeed, not a week goes by where I don't get in front of one large IT organization or another, put the concepts out there, and watch the debate go back and forth between different factions within the IT group.  For those IT groups that end up not liking the concept of a Vblock, that's fine -- we have plenty of piece-part technologies we're more than willing to offer you: storage, backup, security, management, services, etc.

But, as all of us working in this space have noticed, enough progressive IT organizations have embraced the concept (or want to) that we're struggling to keep up with demand.  

That's a high-class problem to have.

How To Build A Direct Competitor To A Vblock

After reviewing the competitive material from HP, Hitachi, NetApp, and a few others, I felt obligated to offer up some suggestions on the areas they'll want to focus on going forward.

First, you've got to think in terms of target use cases -- and who is *not* your customer. 

The base Vblock use case is simple: VMware at considerable scale for a mix of variable enterprise-class workloads.  Doing VMware at scale?   Consider a Vblock.  It doesn't try and be all things for all people.

Second, you'll need best-of-breed technologies.  I know, that's somewhat in the eye of the beholder, but there's not too much debate that VMware, Cisco and EMC are pretty good at what they do.  All three vendors lead their respective markets by a considerable margin.  And any chain is as good as is weakest link.

Third, you'll have to go beyond the ususal core components, and offer both integrated infrastructure management and integrated security management.  Simply waving around a portfolio of potential partners and saying "so many choices!" has proven to be ineffective -- you'll need to invest in creating a baseline of integrated management and security functionality that customers can build off of if they need to.  

Also be prepared to do the same for large-scale backup and DR.  

Since you're selling to serious IT organizations, better have serious IT answers for these topics.  Hand-waving won't do.

Fourth, you'll have to make very hard decisions as to where to limit customer choices, and where to permit them.  Simply  claiming that "everything works with everything in any possible configuration" doesn't really move the needle: the trick is to figure out where hard configuration standards need to be enforced, and where to permit customer flexibility without degrading the overall value  proposition.  

Be prepared for some pretty heated internal discussions.

Fifth, you'll need a considerable engineering organization that's tasked with designing, integrating and validating the integrated results.  Part of the appeal of a Vblock is that it's a tangible product (not a whitepaper), and tangible  products require dedicated product engineering teams behind them.  It isn't a part time project.

Sixth, you'll need an equally considerable solutions organization to stand up supersized helpings of real-world workloads,  characterize their individual and aggregate behaviors, and document the results.  And, since it's a very long list of  potential workloads, best to make this a rather large organization as well.

Seventh, you'll need to be able deliver converged and pre-integrated releases (rather than simply an aggregation of  individual product releases) so customers don't have to live in their current dungeon of dozens of individual components,  each with their own release schedule, each with their detailed list of dependencies.  Plan on a single, integrated and  converged release that looks like a single thing.

Are you with me so far?

We're Just Warming Up ...

Doing the aforementioned gets you a decent converged infrastructure product, but it doesn't get you a go-to-market, does it?

Plan on investing in a substantial and dedicated cross-trained customer support organization that doesn't rely on passing trouble ticket escalations across different corporate and organizational boundaries.  Also plan on managing the integrated support process metrics ruthlessly until you're achieving customer satisfaction numbers that are truly phenomonal.

Add to your plan an investment of dozens (or hundreds!) of unique technical presales people (yours and your partners) that will be very hard to find.  Remember, it's not enough to simply talk about one aspect or another, you need smart technologists who  can talk about the whole thing -- end to end.  Your customers and partners will want to hear about the integrated offering, and not simply get a guided tour of the individual components.

While you're at it, consider investing in several hundred professional services people (again, yours or your partners) that will be equally hard to find.  Yes, you'll need all the usual design/build/operate skills -- but you'll need to especially beef up on specific domains in a virtual world: databases, applications, security, operations, etc.  

And don't forget the small army of business consultants that can help IT leadership "sell" this to the business, and help initiate and manage the change in relationship between IT and the business.

Trying to do this without partners would be ludicrous, so be prepared to overinvest in marshalling your channel partners, your SI partners, your consulting partners, and -- especially -- the growing cadre of service provider partners.  Remember, offering potential customers the potential option of being able to consume a Vblock by either buying one -- or renting a portion from a compatible service provider -- is an attractive option.

Finally, I hope you've got access to a large direct sales organization that is privileged to have a trusted advisor status with your important customers.  You'll be coaching them through some interesting transitions, and they'll usually be dependent on you to get to the other side.

And You Will Be On Your Own Vendor Journey

It's not enough to do a point solution, slap together some nice literature, and move on.  You'll need to invest in a continuing stream of enhancements and evolutions -- a converged roadmap -- that customers will depend on.

From a product roadmap perspective, you'll have to align more and more of your engineering investment around the new converged model, rather than investing in making your products attractive in a stand-alone sense.  That alone should provoke some interesting discussions with your product teams :-)

You'll have to train your sales and marketing organization to focus more on accelerating the journey, rather than simply pushing individual piece parts.  You'll have to resist the temptation to make it all about your shiny technology bits --  although some of those may become more important as the conversation progresses.  

That's a big behavior change for an IT vendor, if you think about  it.

Finally, you -- as a vendor -- will have to grapple with the scary strategic thought that -- yes -- the industry is fundamentally changing, and your game plan for the next few years probably won't look like what it did for the past few.

And, If You're Successful

Fail to execute on any major piece of this while working closely with your key partners, and you'll have something considerably less attractive.  As you go through this list -- is any aspect not important?

You'll probably realize that you will need a new organizational structure to accomplish all of this on an ongoing basis.  

Maybe a new joint venture with an independent management structure and resource alignment, like VCE?  That's some heavy-duty strategic lifting to accomplish -- I speak from direct experience. Contrast CEO level strategy and funding commitment with a powerpoint partnership conjured up by the marketing organizations.

Which are you inclined to invest in -- and which will be around in 3 or 5 years?

And then, finally -- if all of this happens, you'll have eventually something that is reasonably equivalent to what a Vblock does today.  Customers will then potentially have different alternatives that are roughly comparable, and not wildly dissimilar.

But that's not the world we live in today.  

Today, we see partial groupings of piece parts, surrounded by great gobs of happy-talk marketing fluff.  For a recent example, see this.  Or this.  Or perhaps this.

The investment and commitment doesn't really appear to be there, nor does it appear to be coming soon.  I can only imagine how this sort of approach will crumble in front of discerning customers who are serious about their IT.  And I speak from direct and personal experience.

At some point, though, this may change.  

Until then, there's really only one serious enterprise game in town to dramatically accelerate your journey to a private cloud -- and that's a Vblock.

Comments

[nate]

Your headline made me think of 3cV which came out about 3 years ago - (3)PAR storage, HP (c) class blades, and (V)mware.

http://www.3par.com/news_events/press_releases/20071210.html

"Validated Blueprint of 3PAR, HP, and VMware Products Can Halve Costs and Floor Space"

course it was less "official" from HP's and vmware's standpoint I'm sure..

Only thing missing is better networking, I'll take HP Flex10 over cisco any day of the week but would like something even more powerful and easy to use..

It was funny, in the heat of the bidding war between HP and Dell for 3PAR, 3PAR came out an tossed out a press release for another customer signing onto the 3cV bandwagon during VMworld. I know it was coincidental but I can't help but wonder what the Dell execs were thinking when they saw that and knew they had nothing to compete against the c-Class with.

[www.facebook.com/profile.php?id=1545193865]

@Nate... I think the point Chuck is making is that where the true test comes is in sales, services and support execution and, even when all of the components are well engineered, a failure in those areas can break a solution.

History if rife with examples like 3Par and there are a ton today. They are all historical footnotes, for the most part, because, as Chuck points out, they were CEO level strategy with funding and conceptual agreement, but little or no follow through, and metrics alignment, at execution.

AFAIK, VCE is the first time a solution stack has been "legitimized" by the participants at this level of depth in this particular space.

Interesting entry... In the interest of transparency (which I think is a rule on this blog :) ), I recently joined Acadia from "a very large competitor of VMWare", but part of what drove my decision is precisely what Chuck is describing here - the unprecedented amount of skin that Cisco, EMC and VMW are putting into this game.

[Mike Riley]

I would suggest that the integrated stack approach does make sense. Right now, I would say it's more projection than fact that VBlock is "enormously popular" or that the customers are "legion." With 62 customers "beginning Vblock implementations or have Vblock actually in production" and a 56% growth rate (translation: ~40 ordered over the past year with some in a POC and some in production) it's proof of a young and growing idea. Whether or not EMC can barely keep up with production for this demand, I can't say. But, hype can damage adoption - make this idea a target - so, a dose of humility might help us all until we can see how the market votes. We should check back in a year with a better historical perspective.

As far as the rest of your checklist, a NetApp Flexpod hits the mark. If not a Flexpod, you have Cisco validated blueprints that have been implemented (ie. in production) so both consumption models prove to be effective. NetApp has the resources (e.g. sales force, channel strength and PS). Now it comes down to execution. We'll see. I'll keep my powder dry on whether these are "flying off the shelves" though.

[Chuck Hollis]

Hi Mike (of NetApp?)

Go through the checklist I published. Based on the published information from NetApp, few -- if any -- of the boxes can be checked.

Like I've said before -- reference architectures are a dime a dozen these days. A Vblock is a product you can buy and deploy.

It's not a blueprint of a car -- it's a car.

-- Chuck