The Nutanix Acropolis hypervisor streamlines host and cluster management for virtualization administrators.
When people talk about Nutanix Acropolis, they're typically referring to the software used to run VMware and Hyper-V on top of a Nutanix Enterprise Cloud. The Acropolis hypervisor refers to the KVM-based hypervisor stack with next-generation tools and techniques that add value.
The commercial version of the Nutanix software is available as a hyper-converged infrastructure appliance from Nutanix, Dell, Lenovo or IBM, or deployed on other server platforms from Cisco and HP. The software comes in three different versions: Starter, Pro and Ultimate.
The Starter version includes basic software and is good for small deployments, while the Pro version includes more management features and is good for large deployments. The Ultimate version includes the full suite of capabilities and is good for multisite deployments.
The Nutanix Sizer can help you size the hardware -- CPU, RAM and disk -- to see which option best suits your individual needs.
Key hypervisor components
There are a couple unique hypervisor components that help ease management for virtualization administrators.
Each node within the Nutanix infrastructure has a controller -- or management -- VM, which keeps all the hardware and software in check. These controller VMs communicate with each other to manage the clusters, including VMs and distributed storage.
Prism, Nutanix's management control plane, provides the user interface to manage clusters and their infrastructure stack, including the Acropolis hypervisor. Nutanix is different than other hypervisor vendors because every node comes with Prism, which is highly available. Prism is automatically deployed whenever a new node is deployed.
Prism hides complexity -- for example, shared storage and its redundancy -- and presents resources in an easy-to-use manner. For example, Prism hides most of this complexity with Nutanix for vSphere from the get-go. Prism presents storage as one large shared disk, but, behind the scenes, storage is shared across all nodes, which improves resiliency.
This means that there is no single point of failure. For example, if you're using vSphere on Nutanix and lose the node on which the vCenter was placed, it can take several minutes for it to come back. This type of high availability is only available with Nutanix Acropolis hypervisor.
To give a real-life example, with the recent issue of the Meltdown and Spectre vulnerabilities, it took a Nutanix-based cluster less than 30 minutes to upgrade, compared to several hours with a normal cluster. This functionality isn't just limited to the Acropolis hypervisor, but also extends to vSphere.
Prism completes a series of extensive checks before it performs a rolling upgrade across the cluster, and it makes sure all the upgrades -- from firmware to the Nutanix Acropolis hypervisor and more -- are performed successfully before it performs the next one. Prism takes care of everything the virtualization administrator would normally have to do, which saves time and money. This simplicity also extends to the fact that Prism is designed to be easy to use out of the box.
Before using Prism, the virtualization administrator has to set up the first node; all subsequent additions are relatively easy and draw all the necessary data from the existing nodes. If you want more storage or compute resources, just add in another node and the Acropolis hypervisor will configure most of its settings automatically after communicating with the existing controller VMs.
As the virtualization administrator adds more and more nodes to the Nutanix cluster, the underlying storage automatically adjusts to improve performance and resiliency and provide greater scalability. Nutanix also continuously re-evaluates the placement of each and every VM in the interest of boosting performance.
This ease of use extends to disk management. The Nutanix infrastructure extends the capacity of the shared disk as more nodes are added. In addition to this, inline deduplication and compression mean that these nodes have a high capacity for VMs.
Speaking from personal experience, Prism makes the whole management aspect easier. In the virtual infrastructures I've encountered, there tend to be fewer day-to-day issues, and those that do pop up are easy to see and troubleshoot.