Nutanix Cloud Clusters (NC2) on Azure¶

Scope¶

Nutanix Cloud Clusters (NC2) on Microsoft Azure: bare-metal host selection and sizing, Azure networking requirements (VNet architecture, delegated subnets, NAT gateways, Flow Gateway), ExpressRoute and VPN connectivity to on-premises, licensing and pricing models, Prism Central management, disaster recovery patterns (NMST, Leap), migration workflows from on-premises to NC2, Entra ID integration, and Azure-specific constraints and limitations.

Checklist¶

Why This Matters¶

NC2 on Azure extends on-premises Nutanix clusters to Azure bare-metal infrastructure, providing a consistent management experience through Prism Central across hybrid environments. This eliminates the need to refactor applications or retrain operations teams when migrating workloads to Azure. However, Azure's delegated subnet model imposes significant networking constraints -- the requirement for separate VNets per delegated subnet, the prohibition on NSGs for bare-metal subnets, and the IPv4-only limitation create a fundamentally different network architecture than native Azure deployments. Poor planning of the multi-VNet topology and Flow Gateway sizing leads to throughput bottlenecks and connectivity failures that are difficult to remediate post-deployment. The licensing model choice has major cost implications: PAYG pricing for AN36P is $3.744-$4.716/hr per node ($32,770-$41,292/yr), while 3-year reserved instances reduce this by 57%. Organizations with existing Nutanix capacity-based licenses can use BYOL to avoid double-paying. The DR capabilities (NMST with zero-compute or pilot-light patterns) provide cost-effective cloud DR without maintaining always-on infrastructure, but the 30-day trial limitation means DR testing must be carefully planned. AHV is the only supported hypervisor on NC2 Azure -- organizations currently running Nutanix on ESXi must plan for hypervisor conversion before migrating to NC2.

Common Decisions (ADR Triggers)¶

Bare-metal SKU -- AN36P (balanced, broadest region availability, 768 GB RAM) vs AN64 (compute-dense, 1 TB RAM, 128 vCPUs, fewer regions) vs AN36 (legacy, only 2 regions, avoid for new deployments)
Connectivity model -- ExpressRoute (preferred, can share PC VNet) vs Site-to-Site VPN (requires dedicated FGW VNet, no Active/Active support) vs Azure vWAN (transit connectivity, supports NVA inspection)
Overlay networking -- Flow Virtual Networking (abstracts Azure constraints, 500+ subnets from 1 VNet, requires FGW VMs) vs native Azure VNet peering (simpler, limited by 1 delegated subnet per VNet)
FGW deployment mode -- Single-mode (1 FGW, 10-16 Gbps, simpler) vs scale-out mode (2-4 FGW, requires Azure Route Server, higher throughput, BGP session limits)
Licensing -- BYOL (portable CBLs, predictable cost) vs PAYG via Azure Marketplace (operational expense, no commitment) vs Reserved Instances (36-57% discount, commitment required)
DR pattern -- Zero-compute cold DR (NMST snapshots to Blob, cheapest, longest RTO) vs pilot-light warm DR (3-node cluster running, fast RTO for tier-1 apps, higher cost) vs always-on active-active (Synchronous Leap, zero RPO, highest cost)
Hybrid cloud strategy -- NC2 on Azure (consistent Nutanix management across hybrid) vs Azure native VMs (Azure-native tools, no Nutanix licensing) vs VMware Cloud on Azure (VMware ecosystem continuity)
Hypervisor conversion -- Pre-migration conversion to AHV on-premises (Move tool, test before cloud move) vs migrate-and-convert to AHV on NC2 (single step, less pre-testing)

Version Notes¶

Feature	NC2 on Azure GA (2023)	Current (2025-2026)
SKUs	AN36 (2 regions)	AN36, AN36P (16 regions), AN64 (8 regions)
Max cluster size	28 nodes	28 nodes
Flow Virtual Networking	Limited	GA (FGW active-active, scale-out with ARS)
NMST (Multicloud Snapshots)	Preview	GA (Azure Blob, AWS S3 targets)
Nutanix Leap	Async replication	Async, Near-Sync, Synchronous replication
vWAN support	Not available	GA (transit connectivity, Secure vWAN)
NUS (Unified Storage)	Not available	GA (Files/Objects add-on)
Azure Portal deployment	Not available	GA (deploy/manage via Azure Portal, CLI, PowerShell)

Key changes: - AN64 SKU: Newest addition with Intel 8462Y+ 64-core processors, 1 TB RAM, and 38.4 TB all-NVMe storage. Available in 8 regions. Best for compute-dense workloads (databases, analytics) that previously required multiple AN36P nodes. - Flow Virtual Networking: Matured from limited overlay networking to full FGW active-active with scale-out mode. AOS 6.7+ supports 2-4 FGW instances with Azure Route Server for BGP-based routing. Eliminates most Azure delegated subnet constraints for overlay traffic. - NMST: Nutanix Multicloud Snapshot Technology enables cost-effective DR to Azure Blob Storage without maintaining a running NC2 cluster. NMST Engine runs alongside Prism Central and manages snapshot replication schedules. - Azure Portal integration: NC2 clusters can now be provisioned and managed through Azure Portal, CLI, and PowerShell in addition to the NC2 Console and Prism Central, improving integration with Azure governance workflows.

Responsibility Matrix¶

Area	Microsoft	Nutanix	Customer
Bare-metal hardware	Manages	--	--
Underlay network and data/control plane	Manages	--	--
Nutanix software lifecycle (AOS, AHV, PC)	--	Manages	--
Nutanix licenses	--	Manages	Purchases
Azure networking (VNet, NAT, peering, vWAN)	Supports	--	Configures
Flow Gateway and overlay networking	--	Supports	Configures
Workload VMs and applications	--	--	Manages