With Project Beacon, Nutanix Opens Up And Reins In Cloud Infrastructure

Here is a question for you. Which is more proprietary? The Nutanix hyperconverged compute, storage, and networking platform or a cloud like Amazon Web Services or Microsoft Azure?

It was a trick question. They are all proprietary, but oddly enough if you mix them together, you can get something that is at least somewhat less proprietary and more broadly useful — and therefore with a much larger addressable market. And in the case of Nutanix, if it lets go completely of its own cloud storage, compute and networking and wraps its new Nutanix Central management tools and application and data services around native cloud compute, storage, and networking, it has a chance of creating a universal substrate that spans on premises and cloud without forcing customers to run the Nutanix Compute Cloud stack, called NC2, on those clouds.

This is exactly what Nutanix is hoping to do with an effort called Project Beacon, which was announced at its recent .NEXT 2023 user conference. We sat down with Lee Caswell, senior vice president of product and solutions marketing at Nutanix, to find out more about Project Beacon, which is the next and perhaps final expansion and evolution of the Nutanix server-SAN hybrid that debuted way back in 2011 to much fanfare.

Caswell, who came to Nutanix in February 2022, knows his way around virtual compute and storage. He was formerly vice president of product and technical marketing at VMware, NetApp, and Fusion-io and before that co-founder, chief marketing officer, and chief executive officer at hyperconverged infrastructure (HCI) pioneer Pivot3, which was acquired by Quantum in July 2021.

Our job is to always be on the lookout for the next platform, and so we have paid close attention since it uncloaked from stealth mode a dozen years ago. The Nutanix platform, which has had more names than Elizabeth Taylor had husbands over the years, had all of the hallmarks of a technology that could go mainstream in a big way, and had VMware stubbornly refused to do virtual SANs for longer, Nutanix may have eclipsed it. But VMware woke up and protected its market and Nutanix has grown more slowly than it otherwise might have.

But as we have pointed out before, Nutanix and all of the HCI players have, for reasons that are often perplexing, had a hard time going mainstream, by which we mean having hundreds of thousands of customers worldwide. Soon after founding The Next Platform, we took a stab at trying to understand what was holding HCI back despite its obvious benefits for enterprise customers. After watching Nutanix bleed cash for so many years, in September 2021 we pondered why HCI was not more pervasive given the quality of the Acropolis platform, which is another name for the Nutanix Compute Platform also sometimes called Nutanix Cloud Infrastructure. (Nutanix Compute Cloud or NC2 is NCI running on AWS or Azure.) And we also wondered why it was not more profitable after a decade in the market, and thought that maybe being part of the IBM-Red Hat collective could help the Nutanix HCI platform expand to what seemed to be its natural size. We wondered if Citrix Systems and Nutanix should merge, creating a behemoth that could create a big platform to rival those of Microsoft and VMware. We heard rumors about Nutanix potentially being acquired in October 2022 and suggested that only a big cloud would be rich enough to do so. And as 2022 was coming to a close, we just flat out asked if any of the modern storage providers – with Nutanix and Pure Storage being the examples – could actually make money, or even come close to the booming NetApp and EMC in their respective NAS and SAN markets in the 1990s and 2000s.

This Project Beacon twist is as interesting as it is inevitable, and we have seen it before in other platforms. Here is a good example. Just as the Dot Com Boom was beginning to bust, Vern Brownell, who ran infrastructure for Goldman Sachs, founded Egenera and created its BladeFrame platform and Processor Area Network (PAN) virtualization stack. This was way before VMware came on the scene, but just as server virtualization was taking Unix servers by storm. PAN created what we would call bare metal instances that could be virtualized and partitioned to run workloads, much like hardware partitions on Unix systems. Selling this hardware provide more difficult than anticipated, and seven years later the PAN Manager software was broken free of the BladeFrames and sold as a management tool for the blade server infrastructure made by IBM, Dell, Hewlett Packard, Fujitsu, and NEC. By 2013, Egenera was basically kaput, in this case because all of the blade server makers were creating their own management layer and they had no interest in helping their rivals, particularly Cisco Systems and the converged server-network hybrid code-named “California” and known as the Unified Computing System, or UCS, that shook up the server market from 2009 through 2013 before settling down.

Don’t take the wrong meaning from that comparison, with over 23,000 enterprise customers and $1.3 billion in cash in the bank, we are not suggesting that Nutanix is somehow going to fail. It is just difficult to get competitors with their own vested interests to cooperate, even if it is something that customers want.

And we think that enterprise customers operating at scale do want a unified, managed platform that spans on premises and clouds and we do think that existing Nutanix customers will welcome the Project Beacon effort to provide the same management and control for native compute and storage services on AWS and Azure that are provide for the NC2 stack running on premises or on bare metal instances from those two cloud providers.

The shift from virtualized to bare metal instances for running NC2 in the cloud is an important once, since it lowers the cost of the cloud infrastructure and also allows for the Acropolis hypervisor to be used as a single substrate across on premises and cloudy infrastructure. Otherwise, you have to use the native virtualization on the clouds – KVM on AWS and Hyper-V on Azure, or even ESXi on VMware on AWS if you want to do something that will be a very expensive proposition.

Project Beacon is going to be self-funded in a sense because it aims to make the new Nutanix Central cloud management tool span the clouds and on premises application and data services and help customers not only control the whole shebang from one place, but control the exorbitant costs they have on the clouds.

“Cost concerns are coming up because money is no longer free,” Caswell tells The Next Platform. “As the bills have come in, and interest rates are going up, people are now taking a harder look at how they are spending for infrastructure and how they might optimally locate compute and storage. Our latest survey on this workload movement shows that almost all customers have been moving applications across different infrastructure environments, and we expect that to actually increase over time as people start thinking about where’s the best place to locate things given cost, performance, and data sovereignty issues. The issue is a bit like having the kids load the dishwasher after a holiday meal. They just put everything in there, and when you open the door you see that more can fit if you start rearranging things. When it comes to cloud, we are at this optimization stage. It is interesting that Google and Microsoft are talking about optimization, not cost cutting, too.”

The heart of the Project Beacon effort is Nutanix Central itself runs in the cloud as a service, and it is not just an instance of the Prism HCI management tool pried loose from the Acropolis stack, but is rather a brand new piece of software that treats application and data services running on Acropolis, Azure, or AWS as peers. Gartner is calling such tools “digital platform conductor tools,” and we wonder are they thinking train or orchestra? (Probably orchestra.) The point is to have a common view of similar services running on dissimilar end points.

“Today, our relevance with the Nutanix stack is to hundreds or thousands of datacenter and edge end points, with some extension out to the cloud,” says Caswell. “The opportunity of moving up the stack with Project Beacon and Nutanix Central increases our relevance to millions of cloud users.”

Now here is the neat bit, which we are going to come back to and get some real data from Caswell when it is ready for publication. With Nutanix Central running application and data services and running NC2 on the cloud bare metal instances at AWS and Azure, by applying their own virtualization to them and then overcommitting CPUs and share resources across instances, Nutanix thinks it can cut the cost of providing compute and storage for applications in half compared to loading up Nutanix and running it on stock virtualized, T-shirt sized instances on Azure or AWS.

We are going to be talking to Nutanix and getting the information to explain how this works technically and economically. Half is a big deal.

Sign up to our Newsletter

Featuring highlights, analysis, and stories from the week directly from us to your inbox with nothing in between.
Subscribe now

Be the first to comment

Leave a Reply

Your email address will not be published.


This site uses Akismet to reduce spam. Learn how your comment data is processed.