Choosing the Right Infrastructure for Your Enterprise Converge CPU Fight the FUD HCI netapp Nutanix performance Scale Technology User interface

NetApp HCI Versus Nutanix – The Rebuttal


I used to be made conscious of a current article from Rob Klusman at Netapp titled “Netapp HCI Verses Nutanix” by a Nutanix Know-how Champion (NTC) who requested for us to answer the article “trigger there’s some b*llsh*t in it”.

I prefer it when individuals name it like it’s, so right here I’m responding to the bullshit (article).

The primary level I want to handle is the ultimate assertion within the article.

NetApp HCI is the primary selection, and Nutanix is the second selection. Main in an economics battle simply doesn’t work if efficiency is missing.

Rob rightly factors out Nutanix leads the financial battle so kudos for that, however he follows up by implying Nutanix efficiency is missing. Correctly Rob doesn’t present any comply with up which could be discredited, so I’ll simply depart you with these three posts discussing how Nutanix scales efficiency for Single VMs, Monster VMs and Bodily servers from my Scalability, Resiliency & Efficiency weblog collection.

Half Three – Storage Efficiency for a single Digital Machine
Half four – Storage Efficiency for Monster VMs with AHV!
Half 5 – Scaling Storage Efficiency for Bodily Machines

Rob goes on to make the declare:

Nutanix needs infrastructure “islands” to unfold out the workloads

That is simply incorrect and never solely is it incorrect, Nutanix has been recommending combined workload deployments for a few years. Right here is an article I wrote in July 2016 titled “The All-Flash Array (AFA) is Out of date! the place I conclude with the next abstract:

I particularly state combined workloads together with enterprise essential purposes are supported with out creating silos. It’s essential to notice this assertion was made in July 2016 earlier than Netapp had even began delivery (Oct 25th 2017) their Three-tier structure product which they proceed to incorrectly confer with as HCI.

Gartner helps my assertion that the Netapp product isn’t HCI and states:

“NetApp HCI competes instantly towards HCI suppliers, however its answer doesn’t meet Gartner’s practical definition of HCI.”

Combined workloads is nothing new for Nutanix, and never solely is mixing workloads supported, I regularly advocate it because it will increase efficiency and resiliency as described intimately in my weblog collection Nutanix | Scalability, Resiliency & Efficiency.

Now let’s handle the “Key Variations” Netapp declare:

Consumer interface. Each merchandise have an intuitive graphical interface that’s nicely built-in into the hypervisor of selection. However what’s not apparent is that simplicity goes properly past the place you click on. NetApp HCI has probably the most in depth API out there, with integration that permits finish customers to automate even probably the most minute options within the NetApp HCI stack.

The philosophy of Nutanix intuitive GUI (which Netapp concedes) is all options within the GUI have to be made obtainable by way of an API. Within the PRISM GUI Nutanix supplies the “REST API Explorer” (proven under) the place customers can simply perceive the out there operations to automate something they select.



Subsequent up we now have:

Versatile scale. How scaling is completed is essential. NetApp HCI scales in small infrastructure elements (compute, reminiscence, storage) which are all interchangeable. Nutanix requires progress in particular block elements, limiting the alternatives you can also make.

When distributors assault Nutanix, I’m all the time stunned they attempt to assault the scalability capabilities as if something, this is among the strongest areas for Nutanix.

I’ve already referenced my my Scalability, Resiliency & Efficiency weblog collection the place I’m going into loads of element on these subjects however briefly, Nutanix can scale:

  1. Storage Solely by including drives or nodes
  2. Compute Solely by including RAM or nodes
  3. Compute + Storage by including drives and/or nodes

Again in mid 2013 once I joined Nutanix, the declare by Netapp was true as just one node sort (NX-3450) was obtainable, however later that very same yr the 1000 and 6000 collection have been launched giving extra flexibility and issues have continued to grow to be extra versatile through the years.

At present the pliability (or versatility) in scale for Nutanix options is second to none.

Efficiency. At this time, it’s an absolute requirement for HCI to have an all-flash answer. Spinning disks are barely inexpensive, however you’re sacrificing manufacturing workloads. NetApp HCI solely presents an all-flash answer.

Congratulations Netapp, you do all flash, identical to everybody else (however you got here to the celebration years later). There a many use instances for bulk storage capability, be all of it flash or hybrid, Nutanix offers NVMe+SATA-SSD, All SATA-SSD and SATA-SSD+SAS/SATA HDD choices to cowl all use instances and necessities.

Not solely that however Nutanix permits mixing of All Flash and Hybrid nodes to additional keep away from the creation of silos.

Enterprise prepared. This is a vital check. One downfall of Nutanix software program operating on precisely the identical CPU cores as your purposes is the impact on enterprise readiness. Lots of our clients have shifted away from Nutanix as soon as they’ve seen what occurs when a Nutanix element fails. It’s simpler to maneuver the VM workload off the present Nutanix system (the one which’s failing) than it’s to attend for the repair. Nutanix doesn’t run optimally in hardware-degraded conditions. NetApp HCI has no such drawback; it could possibly run at full workloads, full bandwidth, and full velocity whereas any given element has failed.

It’s an enormous declare by Netapp to dispute Nutanix’ enterprise readiness, contemplating we’ve got many extra years of expertise delivery product however hey, Netapp’s article is proving to be with out factual foundation each step of the best way.

The great thing about Nutanix is the power to self heal after failures (hardware or software program) after which tolerate subsequent failures. Nutanix additionally has the power to tolerate a number of concurrent failures together with as much as eight nodes and 48 bodily drives (NVMe/SSD/HDD).

Nutanix can even tolerate a number of failures and FULLY self heal with none hardware being changed. That is essential as I detailed in my submit: Hardware help contracts & why 24×7 four hour onsite ought to not be required.

For extra particulars on these failure situations checkout the Resiliency part of my weblog collection Nutanix | Scalability, Resiliency & Efficiency.

Workload efficiency safety. Nobody ought to try a complicated HCI deployment with out workload efficiency safety. Solely NetApp HCI supplies such a assure, as a result of this safety is constructed into the native know-how.


One crucial think about delivering constant excessive efficiency is knowledge locality. The additional knowledge is from the compute layer, the extra bottlenecks there are to probably influence efficiency.  It’s necessary to Consider Nutanix’ unique & distinctive implementation of Knowledge Locality to know that options resembling QoS for Storage IO are options that are essential with scale up shared storage (a.okay.a SAN/NAS) however when utilizing a extremely distributed scale out structure, noisy neighbour issues are all however eradicated by the very fact you will have extra controllers and that the controllers are native to the VMs.

Storage QoS is added complexity, and solely required when a product comparable to a SAN/NAS has no selection however to cope with the IO blender impact the place sequential IO is acquired as random because of competing workloads, this impact is minimised with Nutanix Distributed Storage Material.

Shared CPU cores. One key technical distinction between the Nutanix product and NetApp HCI is the idea of shared CPU cores. Nutanix has processes operating in the identical cores as your purposes, whereas NetApp HCI doesn’t. There’s a value related to sharing cores when purposes like Oracle and VMware are licensed by core rely. You truly pay extra for these purposes when Nutanix runs their processes in your cores. It’s necessary to try this math.

I’m very joyful Rob raised the purpose relating to VMware’s licensing (a part of what I’d name #vTAX), this is likely one of the many nice causes to maneuver to Nutanix subsequent era hypervisor AHV (Acropolis Hypervisor).

As well as, for workloads like Oracle or SQL the place licensing is a matter, Nutanix presents two options which tackle these points:

  1. Compute Solely Nodes operating AHV
  2. Acropolis Block Providers (ABS) to offer the Nutanix Distributed Storage Material (ADSF) to bodily or digital servers not operating on Nutanix HCI nodes.

However what concerning the Nutanix Controller VM (CVM) itself? It’s assigned vCPUs which share bodily CPU cores with different digital machines.

Sharing Bodily cores is a nasty concept as virtualisation has taught us over a few years. Maintain on, wait, no that’s not it (LOL!), Virtualisation has taught us we will share bodily CPU cores very efficiently even for mission important purposes the place it’s finished appropriately.

Here’s a detailed submit on the subject titled: Value vs Reward for the Nutanix Controller VM (CVM)

Asset fluidity. An necessary a part of the NetApp scale performance is asset fluidity – with the ability to transfer subcomponents of HCI round to totally different purposes, nodes, websites, and continents and to make use of them lengthy past the Three-year depreciation cycle.

That is probably the weakest argument in Netapp’s publish, Nutanix nodes may be eliminated non disruptively from a cluster and added to another cluster together with mixing all flash and hybrid. Model new nodes could be combined with another era of nodes, I commonly type giant clusters utilizing a number of generations of hardware.

Here’s a tweet of mine from 2016 displaying a 22 node cluster with 4 totally different node varieties throughout three generations of hardware (G3 being the unique NX-8150, G4 and G5).

22 node #Nutanix cluster w/ combined node varieties (four differing types) supporting combined workloads and eliminating silos.#HCI #sysadmin

— Josh Odgers (@josh_odgers) December 14, 2016

Knowledge Material. The NetApp Knowledge Material simplifies and integrates knowledge administration throughout clouds and on the premises to speed up digital transformation. To plan an enterprise rollout of HCI, a Knowledge Material is required – and Nutanix has no such factor. NetApp delivers a Knowledge Material that’s constructed for the data-driven world.

I needed to lookup what Netapp imply by “Knowledge Material” because it sounded to me like a nonsense advertising time period, and shock shock I used to be proper. Right here is how Netapp describe “Knowledge Material“.

Knowledge Material is an structure and set of knowledge providers that present constant capabilities throughout a selection of endpoints spanning on-premises and a number of cloud environments.

It’s a fluffy advertising phrase however the identical might simply be argued about Nutanix Distributed Storage Material (ADSF). ADSF is hypervisor agnostic which right away delivers a a number of platform answer (cloud or on premises) together with AWS and Azure (under).


Nutanix can replicate and shield knowledge together with digital machines throughout totally different hardware, clusters, hypervisors and clouds.

So the declare “Nutanix doesn’t have a Knowledge Material” is fairly laughable based mostly on Netapp’s personal description of “Knowledge Material”.

Now the ultimate level:

Selecting the Proper Infrastructure for Your Enterprise

I’ve written about Issues to think about when selecting infrastructure and my conclusion was:


Nutanix has for a few years offered a platform which may be your normal for all workloads and the variety of area of interest workloads that can’t be genuinely supported at the moment are so uncommon with all of the enhancements we’ve made through the years.

One of the best factor about Nutanix, with our world class enterprise architect enablement and Nutanix Platform Skilled (NPX) certification programmes, we guarantee our area S.Es , Architects and authorized people that design and implement options for patrons daily know precisely when to say “No”.

This tradition of buyer success first, gross sales final, comes from our former President Sudheesh Nair who wrote this wonderful article throughout his time at Nutanix

After addressing all of the factors raised by Netapp, it’s straightforward to see that Nutanix has a really full answer because of years of improvement and expertise with enterprise clients and their mission important purposes.

Have you ever learn some other “b*llsh*t” you’d like Nutanix to answer, in that case, don’t hesitate to succeed in out.

About the author