Gartner's HCI "Confusing" Quadrant
Gartner re-created their HCI Magic Quadrant. Here are our thoughts.
Over a year has gone by, and the much-anticipated Gartner Magic Quadrant for HCI has been released. A reprint can be found here if you’d like to dig deeper into the report itself. We reviewed the last report and focused only on the Leaders of the quadrant. Gartner threw a bit of a curveball this year, so we'll cover the quadrant as a whole.
What were they thinking?
Let’s immediately address the elephant in the room. Nutanix and VMware are the only two vendors in the Leaders quadrant. This is because Gartner decided that they should focus on hyper-converged infrastructure software. Notable exclusions from the quadrant are Cisco HyperFlex, Dell VxRail and HPE SimpliVity, as if these solutions don’t require a storage software layer to function.
We read the inclusion criteria, and it appears a single new requirement causes these exclusions: The ability for the vendor to support multiple server brands. Cisco is out because HyperFlex only supports UCS. Dell is out because VxRail (as the solution)* is only supported on Dell servers. HPE SimpliVity is also out because when SimpliVity was acquired, HPE moved to an HPE hardware-only support model.
No real explanation is given as to why this single criterion was introduced (at least not a good one we could find) and why we need yet another quadrant to worry about and confuse customers. We say this because, in the report, Gartner lists out that vendors who were dropped from this quadrant will be included in a future iterated version, HCIS (Hyper-Converged Integrated Systems). Hopefully, a single criterion isn’t the only thing differentiating the two quadrants when more than likely, VMware vSAN and Nutanix will also fit the HCIS quadrant criteria, duplicating efforts.
Is software only enough justification?
Interestingly, Nutanix was cautioned in the previous Gartner report because of its move to software only, yet now this is such an important feature that it requires its own quadrant. Perhaps the requirements for having a new report should be a bit more in-depth. What about the functionality of supporting multi-hypervisors, support for Intel or AMD processors or native cloud extensibility?
Having the choice of server hardware certainly isn’t a feature that should be overlooked by customers evaluating solutions. As a matter of fact, several times we’ve talked to customers who don’t want a certain server brand for a multitude of reasons, which will ultimately lead to the exclusion of some HCI technologies — but not having this flexibility shouldn’t be something that requires a completely new market segment or analysis.
So, is this enough to justify a new quadrant? Gartner specifically calls out that both Innovation and Market Understanding are weighted High as part of the Completeness of Vision score. Supporting multiple server vendors requires an incredible amount of engineering effort to validate components, performance, ease of upgrades for drivers and/or firmware (Innovation), and understanding that many customers want flexibility — multiple server brand support, multi-hypervisor, etc. — can be thought of as market understanding.
The answer: We do not need a new quadrant for software-only HCI. Proper scoring and quadrant adjustments should be made based on these criteria if they are that important.
An SDDC quadrant?
One challenge in this report for Gartner is that they seem to be conflating the definitions for HCI and software-defined data center (SDDC). Gartner’s reference to a “Unified, ‘single pane of glass’ management for virtual compute, storage and networking resources; for the purposes of this analysis, although network management is required, it can be enabled through integrated, third-party software by the HCI software provider” is not actually the task of the hyper-converged infrastructure. Rather, that role is fulfilled by an SDDC, of which HCI is a common component at the compute and storage layers.
This mistaken interpretation of the role of HCI in the modern data center can call into question the overall conclusions of this report. We should think of HCI as an option for providing API-driven storage to an SDDC (or a traditional/hardware-defined data center). That SDDC becomes the API plane for a private cloud to be built upon. If Gartner were to separate these definitions cleanly, it would be a better value proposition for evaluating the role of HCI in an SDDC or a traditional environment.
Fortunately, WWT offers an SDDC briefing that covers this very topic and can help customers properly navigate and evaluate options for building out a proper SDDC.
It's unfortunate that this will more than likely cause confusion with customers, but we're ready to address this and steer everything back in the right direction. We agree that server brand flexibility needs to be a topic of discussion when evaluating different HCI software and capabilities, but not enough to create yet another new market segment.
If you’re serious about evaluating HCI, these are the technologies that should be considered:
- Dell EMC VxRail/vSAN
- Cisco HyperFlex
- HPE SimpliVity
- NetApp HCI
- Dell PowerFlex
Focus on the requirements and the technologies that meet those requirements. Build your own scoring and justification by going through a WWT-led HCI Workshop and walk out confident that the right decision is being made.
* Dell VxRail solution leverages VMware vSAN under the covers but offers tight integration with the hardware.