Solution Design, Sizing, and Pre-Sales Discovery
Discovery inputs
This domain tests whether you can turn workload requirements into a defensible Nutanix design. Gather VM count, vCPU and RAM per VM, used storage and growth rate, IOPS and latency sensitivity, snapshot and backup retention, DR RPO/RTO, and network segmentation. Use Nutanix Collector to profile an existing environment and Sizer to model node count, RF, capacity, and performance. Size for N+1 and include overhead before comparing usable capacity.
Sizing rules
Work in usable capacity after replication. RF2 roughly halves raw capacity; RF3 roughly thirds it. Add overhead for snapshots, metadata, and growth. Check performance per node, not just total capacity. Confirm that the proposed node count preserves N+1 after one node is unavailable. Validate the model against real IOPS and latency requirements.
Worked example
Twenty VMs need 10 TB usable, with 30% growth and RF2 plus N+1. Usable needed is 13 TB. RF2 raw is about 26 TB; adding 20% overhead gives about 31 TB. With four nodes, N+1 leaves three usable nodes, so each node needs about 10.5 TB raw. If each node has 2 x 3.84 TB drives, raw is 7.68 TB per node, or 23.04 TB usable under N+1, which is short. Choose larger disks or five nodes.
Traps
Do not size from raw disk capacity. Do not ignore RF, N+1, snapshots, or growth. Do not size only for capacity while ignoring IOPS and latency. Do not present Sizer output without validating assumptions against the customer's workload.