Your converged / hyper-converged / OpEx offer is not a private cloud

By: Brad Parks

Morpheus has been proud to participate in the GestaltIT Cloud Field Day events for several years.  It’s an unplugged and gloves-off discussion with a group of opinionated customers, partners, and other independent tech gurus about private and hybrid cloud.  The most recent was #CFD17 where we spent a few minutes covering Morpheus 101 and then getting into an overview technical demo with our CTO, a session talking plugin extensibility, and lastly highlighting use cases for self-service automation and IaC.

4 years ago, during my first CFD discussion, my intro hook was a line from The Princess Bride I don’t think that means what you think it means.  I was referring to what passes for private cloud; now years later, it’s inconceivable (get it?) how many IT vendors, IT salespeople, and even IT professionals still mischaracterize what is and is not a cloud. 

Deconverging the converged system landscape

A friend told me the VCE vBlock business started by VMware + Cisco + EMC is finally winding down in 2023 as the last of those systems go end of support.  VCE was one of the first and most successful converged system stacks.  Others followed, using Cisco Compute + VMware Hypervisor + [insert OEM Storage Provider].  NetApp was very successful in this world with the FlexPod, Pure Storage has the FlashStack, and Hitachi has their Adaptive Solutions.  HPE and Dell got into the game with their CS700 and VxRack platforms.

Then we have the hyper-converged set.  LeftHand Networks had one of the first virtual storage appliances but it was Nutanix that really made this market along with the VxRail offer from Dell using VMware VSA.  Simplivity, Pivot3, Hypergrid, and others have also come and gone.  Now, dHCI is the latest incarnation of hyper-converged platforms, with offerings like those from HPE-Alletra.

While all of these converged and hyper-converged offerings each have their own special sauce to differentiate, they all fundamentally solve the same pain – wiring together compute, storage, and VMware in less time with less complexity. 

The other thing that they all have in common is that they do not equal private clouds.  Period.

CapEx, OpEx, and aaS oh my

Over the last few years, as public cloud growth has continued to skyrocket, most of the OEM compute providers have rolled out OpEx-oriented programs HPE GreenLake, Dell APEX, Lenovo TruScale, and Cisco something or other.

The vendors behind these programs have positioned them broadly as clouds to combat workloads going to the public cloud but ultimately the majority of revenue is coming from a more flexible financial model for acquiring and refreshing on-prem hardware. 

Don’t get me wrong; these programs solve a very valid problem and these programs have been quite successful.  CFOs are tired of step-function investments in on-prem gear every 3 to 5 years and the guesswork that goes into capacity planning.  Having the option to commit to a known level but then burst into buffer capacity already deployed on-site and only paying for consumed capacity is a no-brainer. Most also include a managed service component so IT can step away from the day-to-day patching and monitoring of hardware.

But are they clouds? In 2011 NIST defined cloud computing as a model for enabling ubiquitous, convenient, on-demand access to a shared pool of configurable computing resources that can be rapidly provisioned and released with minimal effort.  They highlighted 5 essential characteristics:

  • On-Demand Self Service
  • Broad Network Access
  • Resource Pooling
  • Rapid Elasticity
  • Measured (metered) Service  

Just because your offer is metered, elastic, pooled, and available over a network does not mean you have a private cloud.  Period.

You’re solving the wrong problem

One of my favorite stories in the world of DevOps and design thinking is that of Paul MacCready, a brilliant mechanical engineer who proved the value of solving the right problem when he won a contest to build a human-powered aircraft that could fly over a half-mile.  Check out the full article, but the spoiler is that, unlike others who had tried to win this contest for over 18 years, MacCready rethought the core problem to come up with the right solution which in his case was rapid learning and iteration.

Similarly, for over a decade, on-prem solution providers have been trying to compete with public cloud hyperscalers and along the way most have lost sight of the real problem.  In this case, it’s asking what drove the adoption of public cloud in the first place?  Much of the growth was driven by application developers who needed to test and run their applications somewhere and were tired of waiting for weeks for IT teams to provision the infrastructure they needed. This is still a problem for most large heterogeneous enterprises who have a heavy on-prem or hosted datacenter footprint.

While converged, hyper-converged, dHCI, and OpEx-oriented financing programs solve very valid problems, they are not providing on-demand provisioning of application stacks.  This 5th attribute from the NIST definition is one that has escaped most of those dancing around the private cloud question.   Even Dell Apex Private Cloud has fallen short as it is basically an OpEx-oriented program with VxRail building blocks no self-service app provisioning = no private cloud.

Headed in the right direction

On-demand self-service provisioning is at the heart of Morpheus.  Earlier this year I enlisted my 11-year-old daughter to take a crack at Is it a Cloud? to spotlight the need for this important 5th attribute from the NIST model. (Note to self… next CFD’s movie reference may need to be 1997s Sci-Fi classic The Fifth Element)

With more and more platform engineering and platform operations teams being formed to answer the developer self-service question, we’ve seen our business grow over 50% year-over-year with 6 consecutive quarters of record sales.  While we’re proud of the results, I think this is in large part a reflection of Enterprises moving in our direction and seeking to solve the final piece of the private and hybrid cloud puzzle. 

To their credit, we’ve seen the OEM hardware providers add to their stable of offerings with what we’d consider true private clouds.  HPE GreenLake Private Cloud Enterprise and Dell’s recently introduced Managed Developer Cloud are good examples of combining pre-tested and integrated hardware platforms and managed services with unified orchestration and automation in order to give a self-service app provisioning experience.

No matter how you arrive at the end state, as you think about bringing the public cloud experience on-prem and wrapping governance around your public cloud workloads, Morpheus is here to help. With nearly 1M workloads under management, we’ve learned a thing or two about private and hybrid cloud.  If you’d like to talk to one of our cloud experts reach out and setup a time for a personalized Morpheus demo.

Related Resources

  • card listing image
    Jun 2024 | Blog
    Build vs. Buy? theCube looks at Hybrid Cloud Platform O...
  • listing image
    May 2024 | Case Study
    How infrastructure provider Pulsant provisions, optimiz...
  • listing image
    May 2024 | Solution
    Morpheus + Dell Solution Brief
  • listing image
    Apr 2024 | Solution
    Morpheus Orchestration for Edge Computing
  • listing image
    Mar 2024 | Blog
    NEWS: Hypervisor Independence from Broadcom and VMware
  • listing image
    Mar 2024 | Blog
    Examining VMware license changes in a Broadcom world
  • card listing image
    Feb 2024 | Webinar
    Tech Brief: Deploying an OVA Appliance in Morpheus
  • listing image
    Feb 2024 | Blog
    Reduce risk from the Broadcom acquisition with VMware a...
  • listing image
    Feb 2024 | Blog
    Broadcom-VMware Acquisition:  Aria in trouble with clo...
  • listing image
    Oct 2023 | Paper
    Build vs Buy for Hybrid Cloud Platform Operations