Proxmox, self-hosting and the edge datacenter project

I work on quite a few projects around Proxmox VE and self-hosting. That’s where I test what I can’t break in production — and sometimes the other way around.

Why Proxmox?

Because it’s serious virtualization without the price tag of a commercial hypervisor:

  • VMs (KVM) when you need a full kernel;
  • LXC when a system container is enough — light, fast, efficient;
  • ZFS for storage: snapshots, compression, reliability;
  • Clustering across several nodes for live migrations.

It’s the ideal base for a home lab that can scale up.

What runs at home

A small galaxy of self-hosted services: Git (Gitea), container registries, monitoring, DNS, firewalling — and of course the nodes that double as a playground for the ASN. Everything is versioned, everything is reproducible: I don’t do a task twice if I can automate it, and every personal service is treated like a production service.

From home lab to edge datacenter

The big project is getting this infrastructure to make the leap to an edge datacenter: real colocated hardware, connected to internet exchanges, hosting services with minimal latency and total control.

It’s a multi-step build — but consistent with the rest:

  1. the network first (AS214587, IPv6 prefixes);
  2. the services next;
  3. direct peering on top of transit to finish.

In the meantime, every VM, every container and every snapshot teaches me something — and gives this blog good material.


Written by Rulx Philomé Alexis · ← All notes