Nutanix Fundamentals and Hyperconverged Infrastructure Concepts
What this domain tests
This domain checks whether you can explain the Nutanix model in plain engineering terms: a cluster of x86 nodes where local disks are pooled by software, compute and storage scale together, and resiliency comes from distributed replication rather than a SAN. Know which layer does what: AOS provides distributed storage and cluster services, AHV is the native hypervisor, and Prism is the management UI/API.
Core rules
- Each node contributes CPU, RAM, local SSDs/HDDs, and a Controller VM (CVM). The CVM runs the storage and cluster services; it is infrastructure, not a workload VM.
- There is no external storage array in the data path. The Distributed Storage Fabric (DSF) presents local disks as shared storage to VMs.
- Scale-out means adding a node adds compute, storage, and I/O capability together. Adding disks to one node is not equivalent.
- Resiliency is policy-driven: RF2 keeps two copies, RF3 keeps three. RF2 is normally deployed with at least three nodes so one node can fail.
- Prism Element manages one cluster; Prism Central manages many clusters and adds categories, RBAC, and services such as Flow and Leap.
Worked example
A three-node cluster with RF2 can lose one node and keep data available because each vDisk has a second replica on a different node. Adding a fourth node does not change RF2, but it increases capacity, IOPS, and failure-domain headroom. It also gives the cluster more room to rebuild after a failure. It is not the same as adding disks to one existing node, which would concentrate capacity and risk on a single host.
Traps that cost marks
Do not describe HCI as merely RAID on servers or as only a SAN replacement; Nutanix also virtualizes compute and networking. Do not confuse the CVM with AHV. Do not claim RF2 means two nodes; it means two copies and typically needs three nodes for N+1. Do not assume Prism Central is required to run a cluster.