James Thew - Fotolia

Will Nutanix's KVM-based hypervisor shake up VMware?

Upstart hyper-converged infrastructure vendor Nutanix took a swing at virtualization giant VMware with its own hypervisor offering -- and a way to migrate from ESXi.

The battle between Nutanix and VMware ratcheted up a few notches recently when Nutanix fired a software-defined salvo by announcing Acropolis, the company's free KVM-based hypervisor.

Things have been simmering between Nutanix, the leader in the hyper-converged infrastructure space, and VMware, maker of vSphere and other virtualization products, for quite a while now. The companies are partners; vSphere runs on Nutanix hardware. But ever since VMware unveiled its own hyper-converged infrastructure system dubbed EVO:RAIL at VMworld 2014, things have gotten a tad frosty.

Starting in March, Chuck Hollis, chief strategist in VMware's Storage and Availability Business Unit, launched a few mortars at Nutanix through a concentrated barrage of blog postings. Among other things, Hollis claims similarly equipped Nutanix hardware costs more than EVO:RAIL while delivering inferior performance. Many Nutanix employees responded to dispute these assertions, most notably Lukas Lundell, global director of solutions and performance engineering, who said EVO:RAIL is a flop and, rather than revamp vCenter, VMware has decided to focus its attention on Nutanix.

Then, at its first .NEXT Conference in June, Nutanix fired back by announcing its own KVM-based hypervisor named Acropolis that can be managed with its Prism management platform. Nutanix also revealed it developed a functionality called App Mobility Fabric that allows Nutanix customers to convert hypervisors from ESXi and Hyper-V to Acropolis. Using this technology, a business could reduce -- or outright eliminate -- the need to pay licensing fees to VMware.

SearchVMware asked its advisory board members for their thoughts on these developments and what counter moves VMware may offer.

Sander Van Vugt

Sander Van VugtSander Van Vugt

Nutanix with Acropolis offers a nice option to deliver applications in an easy way, through a complete and open stack.

For those customers that don't have any solution that relates to virtualization, software-defined networking and data centers, it will be an appealing solution. However, this is not what the reality looks like.

Most companies already have an infrastructure in place and are looking for a something that easily integrates with their existing setup -- not something that's going to offer just an alternative.

Click here to see Sander van Vugt's contributor page.

Anthony Poh

Anthony PohAnthony Poh

The day we stop having competition between vendors is the day innovation dies. VMware encroached on the hyper-converged infrastructure market, which Nutanix has a comfortable market share in, so why shouldn't Nutanix spread its wings and enter the hypervisor market with its own flavor of KVM?

Everything is becoming software-defined -- compute, storage, network -- so when every vendor starts offering a similar product, the end customer is bound to become confused unless it really knows what it wants in a solution. That's the problem with too much choice. However, until EULAs are relaxed a little, it's going to be hard to find someone willing to run a fair benchmark between products.

KVM has been playing catch-up in the hypervisor market but recently has gained a lot of traction. One thing it doesn't have is a solid management platform -- something that Nutanix has obviously realized and tried to fix with the launch of Acropolis and Prism.

VMware is still the market leader in virtualization, and many Nutanix customers are running vSphere. Will VMware have anything to fear from Acropolis? Probably not -- in the short term. You could argue that Nutanix is just a distributed file system with a management platform layered on top. However, you can't argue that end users love how simple it is to deploy VMs and applications on a Nutanix system.

The distributed management works well when compared to vCenter Server. The obvious response would be for VMware to do something about its management platform, which frankly hasn't changed much over the years.

As end users start to embrace application-centric solutions, then what Nutanix offers will start to make sense -- the ability to focus on application delivery rather than how the underlying infrastructure will be deployed.

But what's not to say that VMware doesn't have something up its sleeve? We already know that it has created a cloud-native apps division and has started to release products within the DevOps and containers-centric arena with Project Photon and Project Lightwave. It would make sense if it had something down the line that brings all the different products together as a single managed solution.

Nutanix likes to think the future is a turnkey hyper-converged solution where the customer doesn't care about infrastructure. That would be utopia, but in reality everyone still cares about the hardware they deploy the environment on.

VMware is here to stay, but it needs to find some focus. It has a finger in every software-defined pie and has far too many products that come with some kind of fee.

Click here for Anthony Poh's contributor page.

Alastair Cooke

Alastair CookeAlastair Cooke

There is a phony war between Nutanix and VMware, a lot of noise and smoke that signifies nothing. The reality remains that the majority of Nutanix customers use VMware products on their hyper-converged platform.

I'm sure that, over time, this will change. Not as customers convert from vSphere to Acropolis. The change will be the Nutanix Acropolis platform bringing hyper-converged infrastructure to use cases that would never warrant VMware's products. The release of Acropolis is based on the idea that not everyone who can benefit from virtualization wants to pay VMware's price -- or even Microsoft's price.

The Acropolis hypervisor and management platform are an interesting play by Nutanix. I'm sure that a lower-cost hyper-converged platform will enable a lot of use cases that have not been seen before. I can see two types of immediate clients. Service providers and SMB customers will both love having a lower-cost option. Funnily, these two types of customers are at opposite ends of a caring spectrum.

Service providers care deeply about the hypervisor, and how they can integrate into the hypervisor. For service providers, a rich API they can integrate into their existing automation tools is crucial. For service providers, Nutanix will need a great set of APIs that are well documented, easy to integrate and expose as much as possible.

SMB customers desperately want to not care about hypervisors. SMB customers don't have enough staff to have a virtualization specialist, so they will love a platform that is largely invisible. SMB customers don't want to think about the infrastructure. For SMB customers, Nutanix will need to build a simple management interface that hides as much as possible.

Click here for Alastair Cooke's contributor page.

Michael Stump

Michael StumpMichael Stump

Competition is very welcome in the infrastructure virtualization market. I don't think Nutanix needs to make a single sale based on Acropolis to make their point loud and clear: The management components of vSphere have failed to keep pace with customer expectations of ease of use and performance. The stuttered movement from the vSphere desktop client to the Web Client has become a persistent problem, too. Nutanix certainly has a reputation for being bold and brash, and it recognized the opportunity to go head-to-head with VMware. Many of us knew that a Nutanix hypervisor was inevitable, but the quick and easy migration from vSphere to Acropolis caught everyone off guard.

To catch up, VMware needs to make the administration and management of its flagship product fun again. Focus on usability from the customer's perspective. Find out why we still live in the C# client for most of our day, and only flip to the Web client when we have to. And since Nutanix doesn't separate its storage from its hypervisor solutions, maybe it's time for VSAN to be part of vSphere Enterprise Plus.

Virtualization was initially the great abstractor of storage, network and server. Ironically, VMware now offers discrete solutions in each of these segments. We've moved from physical to virtual silos, with storage, server and network now isolated software resources. VMware delivers these resources as products; Nutanix delivers these resources as a solution.

Click here for Michael Stump's contributor page.

Mike Preston

Mike PrestonMike Preston

Although I haven't experienced Acropolis first-hand, I must say what I have read is compelling. Is it a threat to VMware? It depends on what area inside of VMware we are talking about.

As a pure hypervisor comparison, I don't think it is a threat -- not in terms of current customers anyways. I just don't see a ton of VMware shops migrating off of ESXi and running their VMs on the KVM-based hypervisor.

New customers may be a different story. It may all come down to a price and feature comparison, which I've not yet seen. There are other companies such as Scale Computing that do something similar with a KVM-based hypervisor. I think there is definitely a place for these hyper-converged KVM-based all-in-one products in the SMB/ROBO space. Acropolis is certainly a good fit for that.

What VMware should be concerned about is how Acropolis is presenting itself in terms of management. Acropolis takes an application-centric point of view when it comes to management, leaving the VM and the container behind. While there is still lots of room to manage everything at a VM level, Nutanix definitely is on to something; at the end of the day, it's the application that businesses care about.

Click here for Mike Preston's contributor page.

Next Steps

What you should ask before selecting an open source hypervisor

Hyper-converged products are just starting to heat up

What you should know about free hypervisor options

Dig Deeper on Selecting storage and hardware for VMware environments