NCA-GENL key terms

An architecture that bundles compute, storage, and virtualization resources into nodes that are managed as one pooled system, replacing separate servers, storage arrays, and storage networks with software-defined equivalents.

The distributed software layer of the Nutanix platform that runs on every node and provides the clustered storage, virtualization management, and health services that make up the core product.

A service virtual machine on each node that owns the node's local disks and serves I/O for the cluster, presenting storage to hypervisors over the internal network rather than through a hardware controller.

The native hypervisor included with the platform, based on KVM, managed through Prism and designed to work with distributed storage without requiring a separate hypervisor license.

The per-cluster management interface that handles local operations such as creating VMs, storage containers, snapshots, and alerts, and shows the health of a single cluster.

A multi-cluster management appliance that aggregates several Prism Element clusters into one inventory, enabling centralized dashboards, categories, policies, and cross-cluster operations.

A logical grouping of physical disks across cluster nodes from which storage containers are carved; it defines the failure domain and capacity that the distributed storage layer manages.

A logical namespace presented to the hypervisor from a storage pool, holding VM disks and defining policies such as replication factor, compression, deduplication, and advertised capacity.

The number of copies of data the cluster keeps, commonly two or three, so that the system can survive one or two simultaneous component failures without data loss.

A space-efficient data protection scheme that splits data into fragments with parity, using less raw capacity than full replicas while still tolerating disk or node failures.

The background service in AOS that scans stored data, performs garbage collection, checks replica consistency, and drives disk balancing and rebuild work after failures.

The AOS service that handles read and write I/O requests from the hypervisor, coordinates replication to other nodes, and manages the local disk extents holding the data.

The fixed-size unit in which the distributed storage layer writes and replicates data on disk; extents are grouped and tracked by metadata rather than by traditional volume blocks.

The behavior where a VM's read requests are served from the copy of its data held on the same node, reducing network traffic and latency when replicas exist locally.

The process of recreating lost data copies onto remaining healthy nodes after a disk or node failure, restoring the configured protection level over time.

A consistency group of VM disks that are snapshotted together and replicated to a remote cluster as one unit, used for application-consistent disaster recovery.

A point-in-time, space-efficient copy of VM disks or a protection domain that can be used to roll back changes or as the source for replication and cloning.

The maximum amount of data, measured in time, that an organization accepts losing after an outage; it drives how frequently replication or backups must run.

The target duration within which a service must be restored after a disruption, guiding decisions about failover automation, standby capacity, and runbook design.

A disaster recovery method that ships scheduled snapshots to a remote site, so the remote copy lags the source and some recent writes may be lost on failover.

The act of switching workloads or services to a standby site, cluster, or node after a failure so that users can continue working with minimal interruption.

A cluster setting that detects a failed node and automatically restarts its VMs on surviving nodes that have enough reserved capacity to absorb the workload.

The ability of the platform to keep serving data and running workloads despite component failures, achieved through data replicas, HA restarts, and redundant network paths.

Capacity held back on each node in an HA-enabled cluster so that VMs from a failed node can be restarted elsewhere without overcommitting memory or CPU.

A capacity-saving feature that detects identical data blocks or files and stores only one copy, with references pointing to the shared copy; it is most effective on similar data.

A storage efficiency feature that reduces the bytes written to disk by encoding data more compactly, trading some CPU work for lower capacity consumption.

A key-value label applied to VMs, hosts, or other objects in Prism Central so that policies, searches, and reports can target groups of objects consistently.

A policy that keeps specified VMs together on the same host or forces them apart on different hosts, used to match application topology or avoid single points of failure.

The software network bridge on an AHV host that connects VM virtual NICs to physical uplinks and carries both management and guest traffic across VLANs.

An aggregation of two or more physical NICs on a host into one logical link for redundancy or increased throughput, configured with a supported balance mode.

A logical network segmentation applied to switch ports and VM NICs so that traffic from different groups shares physical links while remaining isolated at layer two.

The AOS service responsible for cluster configuration, node discovery, and forming the distributed cluster so that nodes can communicate and share state.

A set of nodes running AOS that present a single pooled system for compute and storage, managed through one Prism Element instance and one cluster VIP.

The hosted services layer that connects on-premises clusters to vendor-run capabilities such as licensing, inventory visibility, health telemetry, and lifecycle recommendations.

A Prism framework that inventories software and firmware components, checks compatibility, and orchestrates upgrades of AOS, AHV, firmware, and related packages.

A pre-upgrade and troubleshooting assessment that inspects configuration, capacity, and component state to flag conditions that could block an upgrade or degrade resiliency.

A notification raised by the platform when a threshold, failure, or policy condition is met, visible in Prism and optionally forwarded to email or external systems.

The web channel where customers open cases, upload logs, review known issues, and download software; cluster serial numbers and entitlement records are used for verification.

The log collection mechanism that gathers diagnostics from CVMs and hosts into a bundle for support analysis or for attachment to a service case.

A workload mobility capability that migrates VMs between clusters or sites with minimal downtime, using replication and cutover to relocate running workloads.