Nutanix Cloud Infrastructure

Three nodes for a standard cluster; one-node and two-node clusters are supported for ROBO/edge (two-node needs a witness).

NCI Starter is capped at 12 nodes per cluster; Pro and Ultimate have no licence cap. AHV clusters are tested to 32 nodes; scale beyond that with more clusters under Prism Central.

Supported for ROBO/edge (CVM needs 6 vCPU / 20 GB minimum); data is protected across disks only and there is no node-level HA, so replicate to another site.

Supported; a Witness VM (2 vCPU, 6 GB RAM, 25 GB) at another site arbitrates failures. AOS 7.6 lets a two-node AHV cluster be expanded to three nodes by adding a node.

AHV Metro Availability: synchronous replication (zero RPO), cross-cluster live migration and witness-driven automatic failover between two clusters within 5 ms RTT. Included in Ultimate, add-on for Pro.

Compute-only and storage-only nodes are supported on AHV clusters; NCI-Compute licences compute clusters that use external storage arrays.

Yes via NCI-Compute: Dell PowerFlex (first, GA 2025), Everpure (Pure Storage) FlashArray over NVMe/TCP (GA Dec 2025) and Dell PowerStore (NCI 7.6, 2026). NetApp ONTAP and Lenovo ThinkSystem storage are announced for later in 2026.

All-flash and all-NVMe nodes supported across NX and OEM platforms; AOS 7.6 scales all-NVMe nodes to 307 TB and up to 40 EDSFF NVMe drives per node. RDMA between CVMs supported on qualified NICs.

Inline and post-process compression plus deduplication, configurable per storage container; included in all NCI editions.

EC-X erasure coding (post-process, on write-cold data) in Pro and Ultimate; not in Starter.

RF2 (one failure) or RF3 (two failures); Starter is RF2 only. Block and rack fault-domain awareness, and a 1 node + 2 disk tolerance option in AOS 7.6.

Redirect-on-write VM snapshots and instant clones with no performance penalty; application-consistent snapshots via Nutanix Guest Tools.

Software data-at-rest encryption (FIPS 140 validated) or self-encrypting drives, with the native Local Key Manager or external KMIP key managers. Pro and Ultimate (included in Pro from 7.3).

Per-VM storage QoS (IOPS/throughput limits) in Pro and Ultimate.

Nutanix Files (SMB and NFS) as part of Nutanix Unified Storage (NUS), licensed per usable TiB; 1 TiB of NUS is included with NCI.

Nutanix Objects (S3-compatible, WORM/object lock) as part of NUS, licensed per usable TiB; NUS 5.3 adds tiering to cloud S3.

Nutanix Volumes presents iSCSI block storage to physical servers and external VMs (e.g. clustered databases); external volume access is in Pro and Ultimate.

Native async replication (1-hour minimum RPO in Starter), NearSync with 1-15 minute RPO (Ultimate, or Pro add-on), one-to-many and cloud targets.

Synchronous replication (RPO 0) between clusters within 5 ms RTT; in Ultimate or the Pro Advanced DR add-on.

Nutanix Disaster Recovery recovery plans in Prism Central: ordered boot, IP mapping, scripts, planned/unplanned failover, failback and non-disruptive test failovers.

Protection policies give scheduled snapshots, local and remote retention and snapshot copies to cloud object stores, but there is no full native backup catalogue; most customers add a partner product.

Secure (immutable) snapshots with multi-party approval for deletion, Objects WORM, and Data Lens ransomware detection and analytics (separate licence).

Integrated AHV backup from Veeam, HYCU, Commvault, Rubrik, Cohesity and others via the Nutanix changed-region-tracking APIs.

Replicate to NC2 clusters on AWS or Azure (or to a service-provider cloud) and fail over with the same recovery plans.

Flow Virtual Networking: overlay VPCs with overlapping IPs, distributed routing, NAT and VPN/BGP gateways, managed from Prism Central (AHV).

Flow Network Security: category-based distributed firewall enforced in AHV, up to 10,000 secured VMs per Prism Central. Included in Ultimate, add-on for Pro.

Granular RBAC in Prism Central, AD/LDAP and SAML 2.0 identity providers, and multi-factor authentication.

Hardened by default to DISA STIG baselines, with SCMA self-healing drift remediation, UEFI secure boot and TPM support.

FIPS 140-3 certificates, Common Criteria EAL2+, DoDIN APL (AOS, AHV, Files), USGv6, and STIGs; ISO 27001 and SOC 2 for cloud services.

Prism Central manages many clusters and sites (on-premises and NC2); NCM 2.0 adds multisite, multidomain management.

Life Cycle Manager (LCM) performs rolling one-click upgrades of AOS, AHV, Prism Central, BIOS, BMC, drive and NIC firmware on NX and supported OEM platforms, including dark-site bundles.

v4 REST APIs with SDKs, Terraform provider, Ansible collection, PowerShell cmdlets and CLIs (acli, ncli). Legacy v0.8/v1/v2/v3 APIs are deprecated in AOS 7.6.

Prism gives health, alerts and performance analytics; capacity runway, right-sizing, anomaly detection and playbooks need NCM Intelligent Operations (NCM Starter or above).

Prism Central projects and quotas; blueprints, marketplace catalogue and cost governance come with NCM Pro/Ultimate (Self-Service, formerly Calm).

AHV live migration, automatic VM HA restart and Acropolis Dynamic Scheduling (ADS) for load balancing; cross-cluster live migration with Advanced DR.

Nutanix Kubernetes Platform (NKP, CNCF-based) runs on AHV, other hypervisors, bare metal (NKP Metal) and public cloud, with Nutanix CSI for Volumes and Files; NKP licensed per core, limited entitlements come with NCI.

NC2 runs the same NCI stack on bare-metal cloud instances in AWS and Azure (also Google Cloud and OVHcloud), managed by the same Prism Central; on-premises licences are portable.

Validated for Citrix, Omnissa Horizon, Azure Virtual Desktop hybrid and Nerdio; NCI-VDI licensing per concurrent user for dedicated VDI clusters.

GPU passthrough and NVIDIA vGPU on AHV in Pro and Ultimate; Nutanix Enterprise AI (model inference and agent gateway) and a full-stack Agentic AI solution (GA planned H2 2026).

Nutanix Database Service (NDB) provisions, clones, patches and protects Oracle, SQL Server, PostgreSQL, MySQL, MariaDB and MongoDB; licensed per core.

Factory-imaged appliances, or Foundation / Foundation Central to image bare servers; a cluster is typically running in hours after racking.

Nutanix NX, OEM appliances (Dell XC Plus, HPE ProLiant DX, Lenovo ThinkAgile HX, Cisco Compute Hyperconverged, Fsas Tech XF), Supermicro and HCL servers, plus public cloud bare metal.

One- and two-node clusters, rugged/short-depth OEM nodes, per-VM NCI-Edge licences (up to 5 nodes, 25 VMs), central Prism Central management and N:1 replication.

Supported for dark sites: LCM dark-site bundles, offline licensing and on-premises Prism Central; Data Lens 2.0 also runs fully on-premises.

Nutanix Move (free to customers) migrates from ESXi, Hyper-V, AWS and Azure with pre-seeding and scheduled cutover; zero-copy in-place migration from vSphere vVols (GA 2026).

Community Edition (CE 2.1) is free for learning and labs on your own hardware (1, 3 or 4 nodes), community forum support only; not for production.

Licences are hardware-agnostic (move between NX and OEM servers) and portable to NC2 in AWS/Azure; Pro or Ultimate required for cloud.

Nutanix supports the full stack on NX; on OEM appliances the OEM or Nutanix is first call with a joint escalation path (e.g. Dell, HPE, Lenovo, Cisco sell and support their own).

Mission Critical: 30-minute P1 response, 24x7x365; Production: 1 hour.

1,200+ validated partner solutions, all major backup vendors, OEM partnerships with Dell, HPE, Lenovo and Cisco, a large community (NEXT) and NCP/NCA certifications.

Licences move across server vendors and clouds, and VMs can be exported, but AHV and Prism are proprietary and leaving requires migration tooling from the target platform.