VIRTUALIZATION ARCHITECTURE

Virtualization is the foundation layer that determines how every workload runs, where it can move, and what it will cost to operate for the next five years.

For two decades, that decision was simple. You bought VMware, you ran ESXi, and the ecosystem built around it. That era ended when Broadcom completed its VMware acquisition and converted a capital infrastructure decision into a recurring subscription event that you cannot renegotiate from a position of strength.

The architects who understand this are not asking “which hypervisor is cheapest.” They are asking which platform fits their workload physics, their team’s operational capability, and their five-year infrastructure trajectory — and then executing that decision before the next renewal cycle forces it for them.

This page maps the full architectural landscape. The Broadcom Exit Strategy is the execution layer for teams already in motion. This is the strategic layer for teams building the decision.

The Hypervisor Is No Longer a Default

Platform lock-in used to be a feature. Broadcom made it visible.

For years, VMware’s dominance was self-reinforcing. The tooling ecosystem, the certifications, the operational muscle memory — all of it created switching costs that kept organisations on vSphere long after alternatives had matured. That was rational. The perpetual licensing model meant the sunk cost was real but the ongoing exposure was bounded.

Broadcom’s acquisition restructured that equation permanently. The transition from perpetual licensing to mandatory VVF and VCF subscription bundles eliminated the stable amortisation model that justified lock-in tolerance. Organisations that expected to carry their existing infrastructure investment for another three to five years are now on a renewal clock controlled by a vendor with no competitive pressure to moderate pricing.

The result is not a vendor dispute. It is an architecture problem. Every organisation still running VMware is now carrying an unmodelled financial exposure that compounds at renewal. The exit is not optional — it is a question of timing and execution quality.

Beyond cost, the Broadcom event accelerated something that was already happening: the maturation of alternatives. Nutanix AHV has closed the enterprise gap. Proxmox VE with Ceph is a genuinely competitive architecture for engineering-led organisations. Microsoft Hyper-V retains relevance for organisations already inside a Microsoft Enterprise Agreement. The market has alternatives that did not exist at enterprise scale five years ago.

→ The Broadcom Exit Strategy: Full Migration Architecture → Post-Broadcom Migration Series

The Physics Every Hypervisor Obeys

Platform changes. Physics doesn’t.

Every hypervisor — regardless of vendor, licensing model, or architecture philosophy — operates within the same physical constraints. Understanding these fundamentals is what separates platform selection from platform gambling.

Every platform optimises these physics differently. AHV embeds the storage and networking layer into the hypervisor fabric. vSphere separates them into licensable components. Proxmox gives you direct kernel access and full control over how each layer is tuned. Hyper-V inherits Windows kernel scheduling behaviour with its own trade-off profile.

The platform you choose determines how much of this tuning is exposed to you — and how much is abstracted away at the cost of flexibility.

→ CPU Ready vs CPU Wait: Why Your Cluster Looks Fine but Feels Slow → Resource Pooling Physics

The Virtualization Stack (Most Comparisons Miss This)

Hypervisors are one layer. The stack is what you actually operate.

Most platform comparisons compare hypervisors. That is the wrong unit of analysis. What you are actually selecting is a complete operational stack — and the hypervisor is only the middle layer of it.

Hardware
    ↓
Hypervisor kernel
    ↓
Storage architecture
    ↓
Networking model
    ↓
Automation / API layer
    ↓
Operations & Day 2 tooling

Every layer in this stack has a cost, a complexity profile, and a dependency on the layers above and below it. When you change the hypervisor, you are not making a single decision — you are making six decisions simultaneously.

The questions that actually determine platform fit:

• Does the storage architecture converge with the hypervisor, or is it a separately licensed and operated layer?
• Does the networking model require a proprietary overlay (NSX, AHV Flow) or can it run on open standards?
• Does the automation layer expose a mature API, or is Day 2 operations a manual process?
• What does the operations tooling cost — in licensing, in engineering time, and in the skills required to run it?

Nutanix AHV is an integrated stack decision. vSphere is a component stack decision. Proxmox is a build-your-own-stack decision. Hyper-V is a Microsoft-ecosystem stack decision. None of these is universally correct. All of them have a workload profile where they are the right answer.

Three Platforms. Three Architectural Philosophies.

There is no universal answer. There is only the right architecture for your workload physics.

Which Workloads Fit Each Platform?

Platform selection starts with workload classification. Everything else follows.

The discipline here is honest classification before platform selection. Most migration failures happen because organizations select a platform for their best workloads and then migrate everything to it — including workloads it was never designed to run.

The Decision Framework

Three questions that eliminate the wrong answers fast.

Migration Reality

A platform decision without a migration model is an incomplete decision.

Choosing a target hypervisor is the first decision. Understanding how you get from your current state to that target — without a production incident — is the decision that determines whether the migration succeeds.

The Broadcom Exit Strategy: Full Migration Physics → Post-Broadcom Series: Technical Execution Layer → Detecting Manual Console Drift: Infrastructure Drift Detection

Frequently Asked Questions

Additional Resources