Dell EMC World 2017: XC Series Updates!

There are a lot of eyeballs on this topic – within Dell EMC, within Nutanix, within our partner community, and most importantly – within our customer base.

So – let me speak plainly: “Dell EMC XC Series and our Nutanix partnership is here to stay”.

Now, I don’t want people to read in too much, or conversely too little into that statement.    There’s a tendency to see ghosts, to presume and assume where no assumption is needed, and of course, there’s no way I (or anyone at Dell EMC) can or will constrain the choices our partners make.

But there are market facts that are material.

• In 2016 Dell EMC’s share of the HCI market skyrocketed, we were the fastest growing player of any significance in the fastest growing segment of IT infrastructure – that’s a fact.
• A large part of that growth came from Dell EMC XC – that’s great. Our XC business is bigger now then it was a year ago, and remains a very material concern for the overall Nutanix business.
• An even larger part of the Dell EMC HCI growth came from the explosion in the rest of the Dell EMC HCI portfolio (VxRail, VxRack FLEX, VxRack SDDC) – that’s great too. VxRail alone grew at north of 208% (and on big numbers – outgrowing the HCI market itself) – ending 2016 at north of a $400M run-rate…. And we’re not slowing down.   Our organic VxRail business is larger than our XC business, and when you add VxRack business it is now materially larger – the majority of our HCI revenues are organic.
• We think winning in the marketplace and being in the strongest position comes from having VxRail, VxRack and XC – and we now have many thousands of customers across the portfolio that depend on us.  

There is a very large re-buy rate amongst our HCI customers – and this is very true of the XC Series business.   Rule #1 for me in business “don’t punch your customers in the face”.

In the first meeting with Nutanix post merger (blog post here – note the consistency) I made a comment to the Nutanix leadership that our strategic posture was simple:

1. We think that for the foreseeable future, the HCI market is not a zero-sum game – that the primary competition is “business as usual”.
2. We think that there are many customers who want a proudly, unabashedly 100% VMware aligned stack.   For them, a singular roadmap and total commitment to be the best answer for a VMware-focused customer is a strong feature.   We are determined to lead that market with HCI offers that take their VMware standard and run with the ball.   This means VxRail for HCI appliances, and if you’re ready for network transformation and NSX, VxRack SDDC for HCI Rack-Scale systems. BUT there are other customers that want HCI that is not deeply bound/deeply integrated with a given stack (XC series for appliances, and if you’re ready for network transformations VxRack FLEX for Rack-Scale systems).
3. We acknowledge that things are a little more funky than before (where for Dell – HCI = XC Series) than now (where for Dell EMC – HCI = Leading Portfolio). I asked: “We are cool with that – are you?”

We continue to put our money where our mouth is – we’re investing into the XC business, even as we continue to grow the rocket ship that is VxRail and VxRack.  

I was listening to a podcast speculating about the end-games for Nutanix.   That’s not for me to comment as a general statement – but I will make one comment: with all that said above, why would we buy them?

And to be clear – while we think that CI and HCI are important, they are really just a foundation for the longer game, the more important game: making the IaaS/PaaS/Data stacks  cloud stacks that run on them simple.  

This is critical for anyone who wants to be the infrastructure leader going forward… and we’re determined that leader is and will be Dell Technologies.

Today’s news is proof positive of continued commitment to Dell EMC XC:

• Dell EMC XC430 Xpress – new lower entry point for the XC Series as low as $25K for a 3 node cluster.
• Integration with the Dell EMC Data Protection portfolio
• Pivotal Cloud Foundry integration for Dell EMC XC using vSphere
• Dell EMC XC Series will be part of Cloud Flex (more to come on that in another post).

Read on for more details!

One thing that is interesting to see in the HCI market is how much the entry price points matter.  An economist would say that HCI as a category has a high price elasticity of demand (PED – more on this here).

Sidebar: this is why its frustrating to see people (including internally to Dell EMC and VMware) who can’t wrap their brains around why there are HCI appliance offers AND HCI Rack-Scale system offers.  You can only shrink the platform and management and orchestration stack of a Rack-Scale system so far.   Example: VMware Cloud Foundation (and VxRack SDDC) are awesome – but it can never go as low as VxRail.   Conversely, if we “stretched VxRail” to cover NSX lifecycle automation and include networking as a standard part of the design – well it couldn’t go low, could it.   Sheesh – don’t be dummies.   Should we have “familiness”?  Of course!   Should we have a simple path for customers?  Of course!   Should there be one thing?  No.  Strategic Rule of Thumb: simplify portfolios as much as you can to cover the market – not one step further.

So - Dell EMC XC can start really small – just like VxRail.    And with Dell EMC XC430 Xpress – we’re taking the XC series lower than it’s ever been before – the headline is “starting as low as $25K for a 3 node cluster”.

Built on the PowerEdge R430 platform – we can get a LOT into a tiny package.

What else are we doing with the XC series?   Like everything we do – we want to leverage the strength of our full portfolio.   This means that we are bringing the strength of our Dell EMC Data Protection portfolio to the XC family – integration with Avamar and Data Domain – as well as the ability to connect XC to public cloud object stores.

Out of the gate – this manifests as a reference architecture, but our goal over time is to approach this the same way we do with VxRail – where data protection, VM-level remote replication, cloud tiering are functions of the appliance itself – built-in, not “bolted on”.

Beyond data protection – there are customers who want HCI to underpin their cloud-native journey.   We have a treat for XC customers looking to do that also…

We’ve done the work to create a Reference Architecture to help customers that want to build a flexible developer platform using Pivotal Cloud Foundry on Dell EMC XC series.

Now, there’s no Cloud Foundry CPI for non-vSphere use cases, so this is for Dell EMC XC platforms that are using vSphere only, but that is the vast majority of our customers.    This is a great step-by-step doc with the best practices approach for building a flexible developer platform, to simplify and accelerate their experiences.   It also includes sizing guidelines for varieties of Application Instance scales and other parameters.

Now – one important sidebar.   I want to make something very clear.   I believe that there will be a continuum of how these HCI stacks are consumed in the marketplace:

1. Software only – the most flexible way, where you can deploy on any hardware layer you want.   You build it, you break it, you own it.   The software will always try to “protect itself” from hardware dependencies, but anyone that thinks/claims that can be totally solved is… silly.   Sounds nice – but silly.   You don’t want to create dependencies on proprietary hardware in a software-defined approach – but even in the open hardware ecosystem – anyone who has been around the block knows that firmware matters, that system level design matters.  It’s not a coincidence that the hyperscale players also design their software and hardware stacks together.   The delusion that you can build a reliable system with a software-only worldview is common among software-only companies.
2. Software + Validated Hardware – a tweak a variant of #1, there’s a subset of an generalized hardware compatibility list that is tuned, sub-selected to be a fit for given software, and then these are packaged together for easier acquisition (not necessarily deployment, but acquistion).   A vSAN-ready node is an example of this.   Nutanix’s efforts for a software-only program and partner ecosystem is an example of this.   It maintains SOME of the flexibility of consumption model #1, but gains only SOME of the outcome orientation of #3.   And that’s intrinsic.
3. Turnkey HCI systems (appliances or rack-scale).   Completely designed as a system – less flexible than #2, and way less flexible than #1 – but if you want a turnkey outcome, the way to go.

Time will tell about which of these 3 models the market prefers.   I’m betting (and it’s not fair, I see the numbers :-) that we get clusters around #1, #3 (which have clear value propositions that are mutually exclusive – flexibility for #1, outcome for #3).   There are massive markets for flexible software-only, and massive markets for turnkey outcomes.

Customers are increasingly realizing the outcome is the name of the game

Let me make it real: my team is responsible for the vSAN ready node program at Dell EMC (and other Ready Nodes like ScaleIO-Ready Nodes, and others) – and I love the Ready Node program, but I’m not interested in trying to make vSAN Ready Nodes approach the value of a VxRail. 

Why?  Because it’s a fool’s errand.  

BTW – a hint – if someone tells you that you can have the flexibility of a “build it yourself” model with the outcome of a “turnkey system” model because of the magic in software (whether that’s AOS, or vSAN 6.6 as examples) – walk away slowly – they are dangerous :-)

The way we approach our HCI appliances is that they are for customers who want a turnkey experience, and that applies up and down the stack.

There’s also some other things that we do that make Dell EMC XC stand out – it’s is not just “Nutanix AOS” + “PowerEdge” = XC Series.  

There is a set of low-level software functions (not called out as a “product” anywhere – but a critical ingredient) called PowerTools.  

PowerTools was developed as part of the learnings of growing the PowerEdge business supporting HCI appliances and underpinning SDS models.    It’s part of the secret sauce that we’re doubling down on – this team (led by a dude named Lewie Newcomb) is my team’s partner within Ashley Gorukhpurwalla’s Server team.   They are the “people interface” and PowerTools is the software interface into the PowerEdge engineering layer as we accelerate our HCI and SDS business – including Ready Nodes.  

We use PowerTools under VxRail, under vSAN Ready Nodes, under ScaleIO Ready Nodes.

What does PowerTools do?  A lot.

These are all things that we discovered you NEED to do if you build systems.   Going forward as we continue to refine our HCI and composable systems management and orchestration layer (VxRack FLEX being another example) – iDRAC, PowerTools and OpenManage essentials are the glue in the server/modular system domain which we are building on.

When you are serious about this HCI business – not just packaging up Nutanix AOS with your server (or taking VMware Cloud Foundation, or Azure Stack and just blaming them on servers for that matter – this isn’t a specific answer, more of generalized architectural comment) and a server – if you don’t think about these things, problems arise.  

That’s why I say: “good luck!” to (customers, technology partners, VARs, you name it) anyone pursuing a software-only approach (whether it’s VMware, or Nutanix, or anyone).   It’s why when people ask me how I feel about Nutanix supporting their own open ecosystem (running on Cisco and HPE for example) I shrug my shoulders.  I don’t really care.  I DO point out that that’s VERY different than our model – which is an OEM, and a singular offer.  You don’t put it together – we do (including all the IP I pointed out above).  We own the whole solution.  

If I were a customer, I would have to wonder how it works when there are very visible public examples where Cisco and HPE poop on the Nutanix model.

Don’t get me wrong – software-only is a valid choice (software only + generalized server = is most flexible path) – but you need to go in eyes wide open – you own putting the stack together ultimately.

When you put all of the above together: 1) new platform that extends the XC series; 2) beginnings of deeper data protection integration; 3) validating PCF on vSphere on XC; 4) continued investment in making sure we deliver turnkey customer experiences, well I’ll say what I said at the start…

So – let me speak plainly: “Dell EMC XC Series and our Nutanix partnership is here to stay”.  

Nutanix may not love the massive growth we’re seeing in VxRail and VxRack, and some may pine for the simpler days when for Dell “HCI = XC” – but I’m sure my Nutanix colleagues do like the growth that Dell EMC XC provides.

We think having the strongest portfolio is good for Dell Technologies, and good for our customers and partners – and that’s what matters at the end of the day.