Gartner's HCI "Vague" Quadrant
In This Article
It's been a few weeks since Gartner has released its HCI Magic Quadrant. We would have released our thoughts more quickly than this but as we were reading the report, we felt the narrative should change a little bit. My peers on the primary storage side of things also wrote a review on the primary storage magic quadrant a few months back. It's a good read and you should check it out, but only after reading this one!
Evaluating Gartner reports
Some of the points in the article published by our peers absolutely apply here as well. A great example of this is the timing between data gathering and report publishing. The lag between the actual state of the market and the evaluation from Gartner can sometimes be significant.
Moreover, the report is updated once a year, at best. At the pace that the HCI landscape changes, it can make this report very inaccurate within a month or two.
Gartner also uses several evaluation criteria methodologies that when we break them down, leaves us wondering why it holds a high weight on overall vendor placement. More importantly, these criteria seem very subjective. A quick example of this is how can we truly score pricing and negotiation of deals. No two deals are ever alike, especially as it comes to pricing and how it was negotiated. Different circumstances for different customers, different competitive landscape, etc. all affect how deals are priced.
Before you keep reading, I want to make sure we are clear: Gartner has a place in the industry, and they do a ton of research which can be valuable to customers and partners alike. Gartner wouldn't hold as much weight as it does if they just threw darts at the wall and looked at what stuck.
The HCI report
There are a lot of vendors included in the report (15 to be exact); some we've honestly never heard of and some we've heard of but see little to no demand. Part of the reason for this is that the requirements to be included in the list from a business aspect are light. Having a minimum of 100 production customers isn't exactly a high bar. For the sake of trying to keep things tight, this article only focuses on the Leaders quadrant, which include the following vendors:
- VMware vSAN
- Dell EMC VxRail
- Cisco HyperFlex
- HPE SimpliVity
Initially, we wanted to take points from the report on the different leading OEMs and expand on what Gartner said in hopes to bring clarity on some of the strengths and cautions that were listed. Maybe next year, I'll point a camera at my face as I take my first pass at the document because I'm sure my reactions were pretty hilarious. A wide range of emotions came over me, from happiness to confusion and frustration, mixed with some actual LOLs (the biggest typed lie on the Internet).
Taken at face value, for the most part, we agree on where the different technologies and vendors land within the leaders quadrant. Paraphrasing what Gartner states towards the end of the report: the decision on which technology to select should be led by business requirements and the technology's ability to meet or exceed those requirements.
Truer words couldn't be spoken. Sometimes, the requirements cannot be met by introducing HCI in the data center, even within use cases that have been proven in the past, but Gartner's positioning of these technologies doesn't give enough details to make sound decisions on what these exceptions could be.
Defining HCI complexities
With all of that said, we do have to thank Gartner for sticking to their guns on definitions of what they consider HCI. One of the functional criteria (among many) vendors must have is to combine VM and SDS resources, both running on the same physical servers, as the primary deployment method. We are in lockstep on that one.
It seems that so many vendors are trying to get on the HCI train by scaling compute and storage as different appliances (disaggregated HCI, dHCI) and then others just randomly drop HCI in their name, often confusing customers with actual capabilities.
We would also point out that based on Gartner's definition, we would have expected to see Dell's FlexOS technology on this list. The product meets all definitions and business criteria listed in the report but isn't there for reasons unknown.
Reading through the document left us with so many more questions than it actually answered, hence why we say it is very vague. As we mentioned before, we know Gartner has a lot of data that makes up these scores, but how they actually scored it is a complete black box. We'd also argue that there are some contradictions within the document that we are unable to wrap our collective heads around:
- What is a strength for one, is a caution for another. Example: One strength found between both Nutanix and VMware vSAN Ready Nodes is their ability to support multiple hardware vendors, and we would absolutely agree! However, for some reason, we only see Nutanix as being "docked" for potential issues with 3rd party support. This doesn't line up, as the customer has a choice (based on OEM) to determine the support offering that makes sense for their business.
- Document contradiction of one product vs. many within a portfolio. Example: Many of the larger OEMs have large portfolios on both the compute and the storage side. Adding HCI and "competing" products within the portfolio gives the OEMs a Caution mark, however, when reading the Context section of the report, it clearly states that "HCI is unlikely to become a full-service platform for IT services across all workloads" and that HCI should be adopted based on use case fit. Yes, absolutely! So why would having a large portfolio be a "caution" when it offers flexibility within the vendor portfolio?
- Platform adoption and cultural resistance. Example: Rightfully so, there are many constraints around enterprises being able to adopt a new hypervisor into an organization, and Nutanix is probably the one most familiar with some of those challenges, so no complaints on that being a Caution. Why, then, is Smart Fabric Services for VxRail a strength? While not impossible, integrating Dell switches with (more than likely) a Cisco data center networking environment adds a lot of complexity and challenges.
- Complexity at scale. Example: VMware vSAN gets a Caution bullet point for the need of careful planning in large scale environments, yet for some reason VxRail doesn't and is literally the same technology under the hood. The point goes beyond those technologies. Any large-scale deployment of any technology should be carefully planned out. Many, if not most, outages in data centers are tied to human error either through lack of planning or lack of operational procedures.
So why does any of this matter? Gartner is a point of reference with a lot of hidden data behind the scenes that isn't available to consumers, which can make it a bit hard to understand context and trust results. If you have no other point of reference, the report is a start, but know there are always caveats and factors that are pulled out of these reports that just don't get into the nitty-gritty details with appropriate weighting that quite honestly, may not even matter to your organization.
If you're looking to get deeper and want to understand what our view is, how we came up with what we did and why, we're happy to go through all of that information with you either through an HCI Workshop or an ATC POC to make sure you're not left to make a decision with incomplete and vague data.