Run applications, not machines.
Proxmox VE and VMware run virtual machines very well. Kubernetes runs containerised services. Apelora Fabric runs an application you drew, with its parts, their connections and their failover, on your own Linux servers.
Why not Proxmox VE or VMware?
If looking after virtual machines yourself is what you want, both are mature choices. VMware is closed source and paid per CPU core. Apelora Fabric is open source, for teams that would rather describe the application and let the platform do the rest.
- You draw the application, not the machines: web servers, a database, a cache, and which connects to which.
- Its network follows the drawing: the application gets its own network, and only the connections you drew are allowed.
- Replicas sit behind one address with health checks, on different servers. A lost replica is replaced, and before a server goes into maintenance its replicas stop receiving requests.
- Risky actions need a second person, and every change is recorded with who made it.
Apelora Fabric runs on Incus, the open source system container and virtual machine manager. It is in early development; see what works today.
Side by side
What each does, by the job you need done. Every statement about another product comes from its own documentation, listed under Sources.
- Yes
- Partly, as written
- No
| What you need | Apelora Fabric | Proxmox VE | VMware | Kubernetes |
|---|---|---|---|---|
| Draw the whole application, then deploy it | Yes: Drawn in Studio | No: One VM or container at a time | Partly: Blueprints: a canvas, finished in code | Partly: Written in YAML files |
| Only the connections you drew are allowed | Yes: Automatic, with OVN networking | Partly: Firewall rules; off by default | Partly: With the vDefend Firewall, a paid add-on | Partly: Network policies, if the network plugin enforces them |
| One address with health checks in front of the replicas | Yes: Built in, with OVN networking | No: Not built in | Partly: With Avi Load Balancer, a paid add-on | Yes: Built in; you add the outside load balancer |
| A lost replica is replaced on another server | Yes: Automatically | Partly: HA restarts the VM; needs 3 servers and shared storage | Partly: vSphere HA restarts the VM on another server | Yes: Automatically |
| Empty a server for maintenance | Yes: Traffic is moved off first | Partly: From the command line; more on its roadmap | Yes: VMs move while running (DRS and vMotion) | Yes: kubectl drain |
| A second person approves risky actions | Yes: Built in | No: Not documented | Yes: Approval policies in VCF Automation | No: Not documented |
| A record of who changed what | Yes: Audit log | Partly: Task logs; an audit log is on its roadmap | Yes: vCenter events name the person | Yes: Audit log, once set up |
| Runs VMs and full Linux systems | Yes: Containers and VMs, through Incus | Yes: VMs and LXC containers | Yes: VMs; Kubernetes through VKS | Partly: Containers; VMs with the KubeVirt add-on |
| Track record | No: Early: tested on a four-server lab | Yes: Stable releases since 2008 | Yes: Mature: ESX 4.0 shipped in 2009 | Yes: Open source since 2014 |
| Licence and price | Open source (AGPL-3.0), free | Open source (AGPL-3.0), free; optional paid subscription | Closed source; paid per CPU core, at least 16 cores per CPU | Open source, free |
Proxmox VE 9.2, VMware Cloud Foundation 9.1 (vSphere, NSX and VCF Automation) and Kubernetes 1.37, as documented on 26 and 27 October 2026. Other tools can add what a column lacks; this compares what each product documents itself.
When to choose another
Choose Proxmox VE when…
You want to look after virtual machines and containers yourself, one by one, with a long track record.
Choose VMware when…
You want an established platform for virtual machines with vendor support, and per-core pricing fits your budget.
Choose Kubernetes when…
Your team builds containerised services, knows the tools, and can add what it leaves to you, such as an outside load balancer.
Choose Apelora Fabric when…
You want to run applications, not machines, on your own Linux servers: draw it once and the network, failover and approvals are handled. It is early software, so start with a lab and talk to us.
Sources
Checked on 26 October 2026; VMware on 27 October 2026. If something here is out of date, tell us and we will correct it.
- Proxmox VE: introduction (history, licence, KVM and LXC), web interface, firewall, software-defined networking, high availability (requirements, failover time, maintenance mode), users and permissions, containers (OCI images), roadmap, subscriptions.
- VMware: VCF 9.1 questions and answers (what is included; load balancer and firewall add-ons licensed separately), licensing model (per core, at least 16 per CPU), designing templates, approval policies, vSphere HA, maintenance mode, audit logging, release history.
- Kubernetes: overview, Dashboard, network policies, Services, readiness probes, ReplicaSets, topology spread, draining a node, RBAC, auditing; KubeVirt.
- Apelora Fabric: what works today, with what was tested on real Incus and what only against the mock.
Proxmox VE, VMware and Kubernetes are trademarks of their respective owners. They are named here only to compare; no affiliation or endorsement is implied.
Tell us about your environment
How many servers, what storage and network, who needs access. We will tell you what Apelora does for that today and what it does not.