4. Cloud Deployment Models¶
Separate from which layers someone manages (Lesson 3) is who else shares the underlying hardware with you.
| Model | Who shares the hardware | Example |
|---|---|---|
| Public | Many unrelated customers, on the provider's shared infrastructure | A typical rented VPS or AWS/Azure account |
| Private | Nobody — dedicated entirely to one organization | pve1 and pve2, owned and run solely for this lab |
| Hybrid | A mix — some private, some public, connected together | Keeping sensitive data on your own servers, while using a public cloud service for everything else |
| Community | A small group of related organizations, sharing dedicated infrastructure among just themselves | Several school districts jointly funding and sharing one data center |
The lab you're sitting in is actually a working example of both public and private cloud in the same course: pve1/pve2 are private (dedicated, instructor-owned hardware), while hermes is really public-cloud infrastructure underneath (a rented VPS on shared provider hardware), even though from your side it just looks like "a server."
Why deployment model matters¶
Where your data physically lives differs by model, and that matters for cost, performance, and compliance — a private cloud guarantees your data never leaves hardware you control, something a public cloud provider can't promise in the same way.
Try it
Classify pve1/pve2 and hermes using the table above, and explain why they're different deployment models even though both are "cloud" in the broad sense.
Tip
"Private cloud" doesn't mean "not really virtualized" — pve1 still runs the exact same Proxmox/KVM virtualization as any public cloud provider. "Private" describes who the hardware belongs to, not how it works underneath.
Play it
Open the Virtualization and Cloud Lab and use the Deployment models tab.