Solution offering
Migrating enterprise VMware at scale
Our solution offering for enterprises leaving VMware ESXi: the virtualisation platform, storage and network foundation, moved as one programme.
It describes how CodiLime runs the migration: how we choose the target platform with you, the six stages and the risk disciplines we apply on every wave, what we add after cutover, and the results of a 10,000+ VM programme we deliver today.
The main topics covered in this white paper:
- The migration service: what we deliver for each layer of a VMware estate
- How we choose the target platform with you, across Proxmox VE, Nutanix AHV, OpenStack, OpenNebula and OpenShift Virtualization
- Our six stages, from assessment to support, and the six risk disciplines on every wave
- What we add after cutover: a self-service Cloud Management Portal, observability, QA automation and data-plane acceleration
- Results from the 10,000+ VM programme we run for a Fortune-500 enterprise, and the engagements behind our approach

Why download this white paper?
How do we choose the target platform with you?
We weigh five criteria, in order: licensing and TCO, SDN and L2 scale, the storage model, proven scale, and operational maturity and APIs. The offering shows how Proxmox VE, Nutanix AHV, OpenStack, OpenNebula and OpenShift Virtualization fit against them.
How do we run the migration?
As one programme in six stages: assessment, proof of concept, platform build, migration, network rework and support. Hypervisor, storage and network move together, with one team for all three.
How do we keep the risk under control?
Six disciplines run on every wave: dependency mapping, a chosen wave model, gated execution with rehearsal, a contained blast radius, verified backups, and a 30 to 90 day hypercare window after cutover. Each produces an artefact you own.
What results has this approach delivered?
On a live programme for a Fortune-500 enterprise: 10,000+ VMs moved from ESXi to Proxmox VE with an EVPN/VXLAN overlay, ~16M network segments in place of ~4,094 VLANs, node restart cut from ~90 to ~15 seconds, and a $0 hypervisor licence.
Related content
- ArticleAutomationONUG Fall 2026 takeaways: Networks must change faster, and few teams have a safe place to test
Krzysztof Wróbel11 min read
- ArticleAutomationWhen build is the right answer: how eBay automated its data center fabrics with Spectron
Benjamin Wharton7 min read
- ArticleAutomationThe complete guide to build vs. buy in network automation
Benjamin Wharton14 min read
- ArticleAI & MLThe LiteLLM Breach: Why This Is an Engineering Problem, Not a Security Problem | CTO Perspective
Krzysztof Wróbel7 min read
Ready to modernize?
Tell us where your data center is today, and we’ll take it from there — all the way to AI-ready.
