Catalyst is networking for people who actually have to run it
I am building Catalyst in the open because networking should be understandable, portable, and owned by the people operating it.
I have spent a lot of time around infrastructure that technically works but is hard to trust.
Not because the engineers are bad.
Usually it is the opposite. Smart people build careful systems, then over time those systems collect exceptions, one-off firewall rules, VPN configs nobody wants to touch, cloud networking assumptions, old docs, and a few "please don't restart that" services.
Eventually the network becomes something everyone depends on and almost nobody wants to change.
That is the problem Catalyst is trying to work on.
Networking should be boring in the right way
I do not mean boring as in limited.
I mean boring as in understandable, predictable, reliable, and you know... boring.
You should be able to answer basic questions without opening five dashboards and asking three people who remember how it was originally set up:
- What systems can talk to each other?
- Why can they talk?
- Where is the source of truth?
- What happens if this node moves?
- Can I rebuild this somewhere else?
A lot of networking tools are powerful, but they make the operator carry too much context in their head. Catalyst is my attempt to build networking infrastructure that feels more direct: clear identity, clear membership, clear routes, clear operations.
Why open source matters here
Networking is too foundational to be a black box.
If a tool is responsible for connecting your servers, apps, users, and customers, you should be able to inspect how it works. You should be able to run it yourself. You should be able to keep using it even if your needs change or the company behind it changes direction.
That is why Catalyst is being built as open source.
Not as a marketing checkbox. Not as "source available, but good luck."
Open source matters because infrastructure is a trust problem. The code, the docs, the deployment model, and the operational assumptions should be visible. If something breaks, you should be able to understand it. If something needs to change, you should not be trapped.
And yes, I would like Catalyst to become a real business.
But I do not want the product itself to be the trap.
The model I care about is hosting, support, advising, consulting, and helping people run it well. If someone wants to self-host it, they should be able to. If someone wants help because they do not want to own every operational detail, I want DeltaOps to be useful there too.
The operator experience matters
In my experience, a lot of infrastructure tools are designed like the operator is an afterthought.
Catalyst is being built from the other direction.
The target user is the person who gets paged. The person migrating a service. The person trying to connect a home lab, a customer environment, a cloud VM, and a small office without turning everything into a pile of special cases.
That means the boring parts matter:
- simple installation
- clear node registration
- predictable certificates and identity
- useful status information
- automation-friendly APIs
- docs that explain what is actually happening
None of that is flashy. All of it matters when you are the one responsible for keeping things online.
What Catalyst is based on
A reasonable question is: what is Catalyst actually built on?
The short answer is that Catalyst is built around Nebula, the open source overlay networking project from Slack.
I fell in love with Nebula years ago after hearing about it on one of my favorite Linux podcasts, Linux Unplugged. It clicked for me immediately. The model made sense. The tool was practical. It solved a real problem without trying to become an entire universe.
The hard part, at least for me, was everything around it.
Managing certificate lifecycle. Getting binaries where they needed to be. Keeping nodes registered. Making the day-two operations feel sane. I came up with a hundred different ways to manage those pieces, and the existing answers never quite fit how my brain worked or how I wanted the workflow to feel.
Catalyst is my attempt to build that missing operational layer around a foundation I already trust.
This is still early
Catalyst is not finished.
That is part of why I want it in the open now. I do not want to disappear for a year and come back with a polished story that hides all the real decisions. I want the project to show its work.
Some parts will change. Some parts have already changed. Some parts are changing right now. Looking at you, Time Crystal. Yes, that is a bad book reference.
Some ideas will get replaced. Some rough edges will be obvious.
That is fine.
The goal is not to pretend infrastructure is clean. The goal is to build something useful enough, understandable enough, and honest enough that other operators can see where it fits, where it does not, and where they might want to help shape it.
What I want Catalyst to become
Long term, I want Catalyst to be a practical open source networking foundation for small teams, homelabs, MSPs, internal platforms, and anyone else who needs private connectivity without turning their environment into a mystery.
I want it to help people connect systems with confidence.
I want it to make ownership clearer.
I want it to be the kind of tool you can come back to six months later and still understand.
That is the bar I care about.
Catalyst is networking infrastructure built in public, for people who actually have to run the network after the diagram is done.