I posted a variation of this on the premium NTNX board, but since I found out about the stock here first, I thought I’d post it here as well:
First, it’s important to understand that the appeal of Nutanix is to replace public cloud usage with Nutanix-run enterprise/private clouds. As Nutanix claims (https://www.nutanix.com/2017/10/09/single-biggest-benefit-en… ) Organizations across the globe are increasingly looking to public cloud – not to reduce infrastructure costs (analyses show cloud is expensive for the predictable workloads that comprise the bulk of applications run in most organizations), but to enhance the ability of the business to react more quickly and effectively to the increasingly dynamic marketplace.
Nutanix provides a way to make enterprise/private clouds as easy to manage as applications in the public clouds, and even be more performant. Nutanix is a software layer on top of existing available software and hardware layers. They call it HCI, which stands for HyperConverged Infrastructure.
Nutanix’s HCI software has two components, which they’ve named Acropolis and Prism. Here’s how Nutanix describes them: Acropolis is a distributed data plane with enterprise storage and virtualization services, and the ability for applications to move seamlessly across hypervisors and in the long run cloud providers. Prism is a distributed management plane that uses advanced data analytics and heuristics to simplify and streamline common workflows, eliminating the need for separate management solutions for servers, storage networks, storage and virtualization.
OK, so maybe you don’t want to play “Buzzword Bingo”. What does that all mean?
From the Acropolis piece you get Distributed Storage. This is what Starrob was attempting to describe. Nutanix replaces the NAS (Network Attached Storage) solutions provided by companies like EMC, HP, NetApp, etc. A NAS is a network that’s dedicated to data storage, independent of the common user network. All nodes on the network have access to all the data in the NAS.
Additionally, Nutanix provides data replication solutions. As the word “replication” implies, this means the data doesn’t just exist in one place, but is copied (replicated) to multiple locations (you can set the number of copies to 2 or 3). This is a safety feature in case of anything from a single drive failure to failure of an entire node. Nutanix also provides for various backup solutions (sometimes known as DR, or Disaster Recover) in which data is copied off of the cluster completely, typically to a different physical location, which may be another cluster or to a public cloud (what they call Cloud Connect. And, as is standard today, to reduce having to purchase 2 or 3 times as much storage as you need, Nutanix supports deduplication and compression technologies.
On top of replication and backup, Nutanix also provides performance advantages for data storage. In what they call Intelligent Tiering, data is placed in an SSD (flash)or HDD (physical spinning drive) tier accordingly, and automatically moves data between the tiers based on usage. I don’t know the details of their algorithm, but typically frequently and/or randomly accessed data would be placed on the faster SSD tier if possible, while large datasets and/or sequentially accessed data would be placed on HDDs.
Finally, in what they call Data Locality, Nutanix attempts to place data on the storage that’s directly attached to the node on which the application is running. This reduces the amount of data that’s transmitted over the network, so is often faster than a traditional NAS, in which all data goes over a network. This is possible because instead of isolating storage onto a separate network through which all data must pass (the NAS solution), typical hardware configurations are that storage remains directly attached to each node. Nutanix tries to have the data for an application running on a node be available directly on that node, but if not, that data is seamlessly provided over the network so the application doesn’t care. This is done by having a Nutanix Controller process (CVM) running on each node in the cluster. These Controller processes talk to each other, so if one node goes down the others step up. And remember, if a node goes down that storage also goes down but thanks to replication that data is still available on the other nodes.
There’s also Shadow Clones in which copied data is read accessible from different VMs or nodes to improve performance. This post is already too long so I won’t get into the details, but having more copies of data increases the odds of it being accessed via direct, local, storage.
This kind of Distributed Storage is nice, but Nutanix is more than just a better file system. Acropolis also contains support for Hypervisors. You may not know what that is. A Hypervisor is software that separates a computer’s Operating System (“OS”) from the underlying physical hardware. Instances of a running OS are called Virtual Machines (VMs), and many VMs can be running at once on a piece of hardware, or on a cluster of nodes (hardware). VMs do their job so convincingly that Applications running in them don’t know the hypervisor exists.
Like people living in The Matrix, an application running inside a VM thinks it’s running directly on real hardware, but it’s really running in a simulation. Yes, at some point there’s real hardware involved (just as people in The Matrix actually have bodies somewhere in the real world), but the real hardware may not look at all like the hardware the applications think it is. VMWare is one company that helps make VMs easy to setup, deploy, and manage for enterprises. They had a great business for a while, but that’s old hat today.
One thing that’s nice about Nutanix’s Acropolis is that it supports not only VMWare’s hypervisor (called vSphere), but also Microsoft’s version, called Hyper-V. Or if you want, Nutanix has its own “free” hypervisor (built on top of the open source Linux KVM hypervisor), which they claim has some advantages, especially around taking advantage of their Distributed Storage solutions, described above. I imagine some customers have already standardized on VMWare or MS virtualization solutions, so it’s good that Nutanix doesn’t require them to change to get onto their platform.
Even more, Nutanix leverages VMs to enable applications and data to be migrated not only between different Nutanix clusters, but from non-Nutanix systems or to/from a public cloud (eg AWS, Azure). You can move within Nutanix systems running different hypervisor implementations as well. Even running VMs can be done “live.”
OK, that’s cool and all, but savvy companies today can piece together a solution involving smart NAS and VMs that’s probably almost as good. But, there are two additional aspects/capabilities that are really compelling.
First is that Nutanix is hardware independent. Sure, you can buy Nutanix hardware for your nodes, but you can also license the software to run on your own hardware.
Second, is how you manage a Nutanix cluster, or system of clusters. This is with their Prism software. As Nutanix says: Prism combines multiple aspects of datacenter management into a single consumer-grade product that lets IT admins manage infrastructure and virtualization, gain access to operational insights, and fix problems all with a few clicks. Just as Acropolis creates a data plane that spans the entire cluster, Prism creates a cluster-wide management plane.
OK, to translate: “consumer-grade product” means that the UI is really intuitive. Not CLIs (Command Line Interfaces), but a true graphical point and click UI with summary dashboards, alerts, and automation. Things like software upgrades are one-click operations that get deployed automatically across the network without downtime (thanks to the redundancies described above). And since it’s all Nutanix, you don’t have separate data upgrades and VM upgrades, etc. There’s some AI in there, with predictive capabilities. In addition, they provide a program called “Prism Central” that helps you manage multiple clusters.
This is big. As I pointed out earlier, Nutanix believes that people turn to AWS or Azure not for price reasons, but for ease of deployment reasons. Setting up an AWS app is easy. Scaling in AWS is easy. Apps can even scale up and down based on demand. And that’s without you buying any hardware, but also, and this is the most important reason according to Nutanix, without you needing to manage it. What Nutanix provides is an easy way to have the advantages of a private cloud (performance and security and cost), with the ease of cloud management.
Where Nutanix is going next is integration with public cloud solutions. They call this Nutanix Xi Cloud Services. Today you can’t instantly run an application on either AWS or your private cloud. You can migrate applications between your Nutanix private cloud and the public clouds, and you can backup your Nutanix cloud to the public cloud. The first available Xi Cloud Service will enable Nutanix customers to set up, manage and test a complete cloud-based DR service in just minutes. Using the same Prism management interface, Nutanix customers can instantly protect their applications and data inline with existing workflows as part of their routine IT operations, avoiding the expense and complexity of a separate DR solution. https://www.nutanix.com/press-releases/2017/06/28/nutanix-re…
Finally, I think one needs to be careful here not to overstate what Nutanix actually does today. Nutanix has some very cool stuff, but they’re promising more in the future in terms of the “hybrid cloud” than they’re actually delivering today. That may mean the company has a big runway of growth, but there’s also risk in if and when they can actually deliver.
Hope this helps explain what Nutanix actually does. You-all can discuss the financials better than I.