First – I want to repeat the headline, because a headline this good shouldn’t be buried: Today, a turnkey Enterprise Hybrid Cloud got 3x easier, 3x smaller, and has an entry price point 5x lower.
This makes the number of customers for which it’s an interesting option for 10x more customers – doubling the market.
Now – let’s look at context and detail. Hint - always focus on cause, not effect when you want to understand the big picture.
For years now, technology leaders never have woken up on a bright shiny morning in IT and said “you know what I need – I need a new server”… or “new storage” ... or “a new network switch will change the game”.
It’s not that there aren’t people that focus on those questions, and for them it’s their world. It’s not that component level IT topics don’t exist, aren’t filled with innovation and don’t matter. Component/ingredient level topics matter – but they are the “effect” not the “cause”.
You may be thinking that I’m leading to a point: “that it’s all about Converged Infrastructure (CI) or Hyper-Converged Infrastructure (HCI)” Nope.
CI and HCI matter more on the big stage than components – but they remain a simplification of “effect”. CI and HCI are a force that simplifies, collapses and automates the common domains of server/network/compute. Don’t get me wrong, CI and HCI are awesome – but they are just a part of the picture.
At their best, CI and HCI represent the foundation of “transformation of IT” (customers getting out of the server/network/storage business), but not a full vehicle for “business transformation”. That said – a strong foundation is essential, and the simplifying power of CI/HCI (in particular HCI) – it plays an important part in what really matters – the “prime mover”
So what is the “cause”, this “prime mover” I’m talking about?
The cause has been and will continue to be a simple idea that should never be lost: IT’s job is to support business applications and critical workloads, and more generally provide platforms for the business via SaaS, PaaS, and IaaS. This is the cause for IT. This is the purpose of IT – it’s “raison d’etre”.
Never lose sight of that. IT’s “cause” is to provide these platforms that enable the business to run and innovate. It’s important for all IT practitioners and innovators to always keep it in mind. Remember cause leads effect, not the other way around. Put otherwise – don’t start from the bottom up, start top down.
This is why one of the most important missions at Dell EMC and ultimately at Dell Technologies isn’t at the component level (though that matters), or even at the CI/HCI level (though that matters) – but how we make turn-key business platforms for IaaS, PaaS – the Multiple, Hybrid Cloud Platforms that are the “cause”.
Let’s break down the key words there: “Hybrid”, “Multiple”:
- The “Hybrid” vs. “Public Only” or “Private Only” debate is really over – except in the minds of a few – particularly those who like hyperbole :-) With notable exceptions to be sure – the answer will be Hybrid – why?
- For data gravity reasons. Compute tends to live where the data lives – and will bias to where the data is born (either on or off premises). This is due to physics, not alt-facts. Speed of light. The economic curves of WAN relative to persistence, compute and local networks.
- For governance reasons. This is NOT thesame as not the same as “security”. I’m tending to ignore people who FUD public cloud security which can be demonstrated as better than many private clouds.
- For economic reasons. Public will always win for workloads that are highly variable, relatively transient – public clouds have an overwhelming advantage for highly elastic workloads, particularly ones that have a compute bias. Conversely, there are fundamental reasons why we are even seeing some repatriation of workloads in private clouds (whether it’s Uber, Dropbox, or others) at the same time that here is an indisputable massive growth public clouds – this is not an OR, it’s an AND.
There is no question that the massive web-scale public clouds have changed the game, and the pendulum is swinging – but the pendulum will settle with a blend of on-and off-premises platforms. This Evaluator Group study shows the simple economic picture of on and off-premises models – not that on-premises always wins either, but rather that with workload variation, it’s both, not one or the other. Net? The answer is “hybrid” – the question has shifted to “what stacks, what workloads, what models” – get used to it.
- If “hybrid” operating models is the answer – you then ask a simple followup question: Does anyone really believe that they will have just ONE IaaS/PaaS/SaaS in every customer? Of course not! Therefore “multiple” becomes critical. Technologies that link/bind together multiple clouds become important. What are examples of “binding” technologies? I’m increasingly skeptical about Cloud Management Platforms (CMPs) that homogenize together multiple IaaS platforms. Yes, I see customers using heterogenous CMPs to do this (and it’s not that “pick your example” are bad in any way) – but rather they homogenize and neutralize all the IaaS layers - without a ton of additional value. What I tend to see as the strategy that seems to win more often is to picking the CMP that is linked to your stack of “most use”, and then binding and extending what you can. For example people use vRealize to manage AWS, Azure and other OpenStack based IaaS. It’s notable that even that has somewhat limited value – because these CMPs all neutralize a given layer (like an IaaS) without a ton of benefit, it’s just that the “pick one” strategy at least has advantages to ONE of the stacks. Conversely, I’m increasingly confident about things that bind without “homogenizing that particular layer”, but more are about “value on TOP of that layer”.
- Examples include: Pivotal Cloud Foundry creating a common PaaS across multiple IaaS platforms
- ServiceNow as an ITSM that spans and binds many services – not by homogenizing them, but by wrapping them in necessary workflows.
- There are examples “inside” the IaaS layer – but only where they provide some big incremental value – an example of that would be NSX and other cross-cloud networking, encryption and security services.
- Heck, there’s an argument that configuration/automation tools (Puppet, Chef, Ansible and the like) are things that bind together multiple clouds, but without the “homogenization” of the CMPs.
Now, this brings us to today’s news – we are now offering Enterprise Hybrid Cloud on VxRail.
Why does this matter? As much as point #1 (hybrid) and point #2 (multiple) above are true – there’s something that we simply must face as an industry:
Currently, deploying, managing, supporting, and all “day 2” lifecycle operations for private clouds is WAY, WAY too hard, and starts WAY, WAY too big (in hardware, and in software complexity).
This is an industry problem for us to solve – and we’re fighting to lead the way at Dell Technologies.
We’ve been working on the Enterprise Hybrid Cloud (EHC) for 4 years now – and have been working with customers of every size, every part of the world. We’ve found a common set of needs that are way beyond a basic IaaS. It’s an IaaS stack engineered for the needs of an enterprise. What does that mean? It’s built around the VMware IaaS (vRealize, NSX, vSphere, and now vSAN in VxRail) – but also includes things that we’ve found most enterprises expect:
- Sophisticated workflows for integrated data protection and DR
- Platform hardening and additional security measures and encryption capabilities
- Multi-site topologies where workflows follow workloads as they move between sites.
The Enterprise Hybrid Cloud is not just IaaS – it’s ITaaS.
Furthermore – together Dell EMC and VMware standup and take single call responsibility for the full stack, not only getting it running, but full lifecycle (patch, upgrade, decommission) and single call support.
EHC has been deployed at some of the largest customers in the world – and deploying hybrid cloud has not been easy – even with great tech and great people. I hate to be so blunt, but we have hundreds of people that have been working on this for years. I cringe to think of a customer taking on the challenge themselves. It’s also historically only available on massive VxBlocks – so between the software, the services, the hardware – the price tag is enormous. Frankly, the customers like it – but it doesn’t feel too “cloud like”.
This is an industry call – if we can’t make ITaaS simpler, easier – then it’s on us.
- So – we have made EHC much simpler. 3x simpler.
- So – we have made EHC much smaller. 3x smaller to start, and with a 5x lower entry pricepoint – bringing it to customers who previously couldn’t afford it.
…But we managed to do it without losing EHC ITaaS capabilities and its ability to scale.
With the release of EHC 4.1.1 on VxRail, we’ve taken a huge step together with VMware.
By pivoting to a strong focus on HCI as a simplifying factor – and using VxRail Appliances, the industry’s best HCI Appliance for VMware, we can start small and grow. There is MASSIVE VxRail momentum. With 8000+ nodes out there, with 100,000+ CPU cores, 65PB of storage – all in 70+ countries in the world, VxRail has been a massive success. Now VxRail also is a “start small” foundation option for the Enterprise Hybrid Cloud.
Just as importantly, we’ve been working overtime to automate most the deployment tasks - EHC is now 3x simpler. EHC is now a great answer for tens of thousands of customers that previously would have seen it as “out of reach”.
EHC can federate to multiple public clouds, including Amazon Web Services. Today we support Azure for non-prod use cases (EHC 4.1.1 uses vRealize Automation 7.1). Soon we will update the stack again, getting to synchronicity with the VMware Valided Design (VVD Program) and vRA 7.x – at which point, Azure, Virtustream and other public clouds will be integrated.
We’re not stopping here. What’s next for the team on EHC across Dell EMC and VMware?
- Better synchronicity. Leveraging the VMware Validated Design (VVDs) program, we’re working together with VMware to make EHC build on that VVD program to stay in lockstep with VMware’s release train for all the SDDC software titles – no small feat. Right now EHC uses vRealize 7.1 – but will move to vRealize 7.3 quickly as an example – and we’re using the VVD process to hammer out the synchronicity of the releases – particularly around NSX and vRealize – two fast moving trains. It’s notable that VVDs, VMware Cloud Foundation (VCF) and EHC will always lag point product releases a little bit – but the goal is not by much.
- Better automation. The automation of the lifecycle is key to deliver a real cloud experience. With the arrival VMware Cloud Foundation (VCF) – many of the tasks the EHC team has slaved over for years – particularly those associated with NSX deployment and life-cycle are handled by VCF. VxRack SDDC is the Dell Technologies manifestation of VCF on hardware – and people can expect the full EHC on VxRack SDDC to come soon.
- Better off-premises packaging. Today EHC federates with public clouds that customers procure. Later in 2017 with the availability of the VMware on AWS offering, we will price and package them together – for a simple on- and off-premises answer.
Observe this – as we merge VCF, VxRack SDDC and EHC – it becomes hugely simplified: you want VMware’s SDDC stack in piece parts on and off-premises, you assemble? Great, buy from VMware, hand assemble. Alternatively, are you more of an “outcome” type? Do you want to be out of the ingredient business and have a turnkey enterprise-ready IaaS stack on and off-premises? Great, buy EHC on VxRack SDDC which will include VMware on AWS options.
The above are all roadmapped, scheduled, and teams are cranking on them. That said, there are things we know we need to do above and beyond the above.
- Get even simpler and smaller. We also know we can go even simpler, even smaller than we have today – but not without pulling out some pieces. There are customers who would be absolutely fine with a subset of the full EHC services catalog, a subset of the vRealize and NSX capabilities – and we will build a solution for them on VxRail.
- Better flexible pricing. EHC 4.1.1 still has a core “capital acquisition/ELA” financial construct, with as simple subscription for the lifecycle activities. “Full Consumption Economic models” are right around the corner – imagine all of the above with a complete “pay as you go” model – with no capital cost. You’ll see this in weeks on HCI, and then we will extend the model to Hybrid Clouds built on HCI.
There’s another “meta” observation I want to give my dear readers – which include a lot of customers, a lot of partners (heck a lot of customers), and I use Virtual Geek as a way to “telegraph” where we are going.
If you read the above, you can infer that we’ve come to an important set of conclusions: 1) go faster on simplification; 2) move fast on new economic constructs; 3) the only way to make clouds easy enough is to use HCI as the foundation. #3 is linked to the ideas I presented in the SDS/HCI blog post series here – and why the statement of “SDS/HCI is ready for the majority of x86 workloads by volume” is really important. In the long game, EHC will ONLY be available on HCI. This will TAKE TIME – so VxBlock customers can breathe a big sigh of relief, but you can see our thinking about the important role of HCI as the foundation for well-run, ultra automated hybrid clouds (VxRail and VxRack SDDC for VMware; XC and VxRack Azure Stack for the Microsoft ecosystem, our Redhat OpenStack Platform 10 validated system also)
I want to offer my congratulations to the Enterprise Hybrid Cloud team – hundreds of passionate folks at Dell EMC and VMware working to make IT and Business Transformation easy.
Cause? Customers need simple, cost-effective Hybrid Clouds.
Effect? Enterprise Hybrid Cloud is now on VxRail.
Are you an EHC or VxRail customer? How is it going – give us your thoughts and feedback!