HCI & Virtualization · VMware Exit · Private Cloud
Licensing shocks turned virtualization from a solved problem back into a board-level decision. I assess your estate, pick the platform the numbers actually support — Nutanix AHV, Proxmox VE, right-sized vSphere, or cloud — and build the cluster: from a three-node starter to a full production estate, migration included. Systems that hold.
hci & virtualization
The same productized ladder as every engagement here: a fixed-price assessment you can act on without me, a build sprint if the numbers say move, and a retainer if you want the day-2 burden carried. No reselling margin, no vendor quota — the recommendation is the one the licensing math supports.
A fixed-price teardown of your current estate: workload and dependency inventory, licensing-exposure math on renewal, hardware compatibility against candidate platforms, and a scored recommendation — Nutanix AHV, Proxmox VE, right-sized vSphere, or cloud — with a sequenced migration plan you own.
fixed price · scored report · 2–3 weeksDesign and build of the production cluster — from a three-node starter cluster (the canonical HCI entry point) to a full rack: network and storage-policy design, node imaging and cluster formation, backup and DR wiring, staged migration waves, cutover, and runbooks your team can operate.
sprint · working cluster · migration includedDay-2 is where clusters live or die: lifecycle and firmware upgrades, capacity engineering ahead of the growth curve, backup verification and DR drills, and a principal on the escalation path when something pages at 2am.
retainer · upgrades · dr drills · escalationplatform coverage
I don't resell hardware or licenses, so the assessment lands where your workloads, hardware, and renewal math point — not where a partner quota does.
Hyperconverged clusters on certified hardware: Foundation imaging, Prism management, replication-factor storage design, CVM sizing (the overhead everyone under-budgets), and lifecycle-manager upgrade discipline. Procurement guidance through the right reseller so your support chain is clean.
3-node RF2 starter → scale-outThe open-source exit ramp: clustered KVM with HA, Ceph or ZFS storage, integrated deduplicated backup, and no per-core licensing. Ideal for cost-driven exits and data-sovereign organizations — with honest guidance on the support model trade-off.
no per-core tax · subscription optionalStaying can be the right answer. I model the renewal against a consolidation pass — host counts, core counts, edition tiers — so if you keep VMware, you keep exactly as much of it as you need.
renewal math · consolidation · leverageThe part that actually breaks: dependency mapping, migration-wave design, storage and network translation, rollback gates per wave, and a cutover rehearsal before the real one. Downtime windows are negotiated in minutes, not hoped for.
waves · rollback gates · rehearsed cutoverproof, not slideware
My own holding company operates an on-prem virtualization estate — enterprise rack servers with out-of-band management, a KVM-family hypervisor cluster, a dedicated deduplicated backup server with verified restores, 10-gigabit storage networking, and automated graceful-shutdown tooling for power events. The disciplines proposed to clients are the ones that page me first.
Enterprise rack servers provisioned over out-of-band management (IPMI), with scripted, repeatable node builds.
Dedicated backup server with deduplicated datastores, scheduled verification, and restore drills — because a backup is a rumor until it's restored.
Help desk to principal SRE: storage arrays, tape libraries, VLAN design, firewall policy, and the incident scar tissue that comes with them.
LLM-driven runbooks, log watchers, and nightly audits keep my fleet honest — the same tooling can watch yours.
One working session. Bring your host inventory and your last licensing quote — leave with a straight read on whether to stay, consolidate, or move, and what the first wave looks like.
consult@chrisbergeron.com