[jbverschoor]

Yup.. no focus. The product is really good. Could use a better interface.

But he should be working on getting it into people's hands, instead of creating new tech.

Seeing this leads me to believe that zerotier wants to do this whole global network thing. This is not what I want from zt.

This probably came from questions like "how can we trust you as a network controller". And the answer is to create some tech. But in that case, why not simply offer an AMI on aws for a controller?

All assumptions of course. Upon further checking:

From https://news.ycombinator.com/item?id=20329060 and https://www.zerotier.com/lf-announcement/ :

> We are going to use it to fully decentralize ZeroTier and for some specific enterprise customer projects, but others can use it for other stuff.

> Our interest in decentralizing the roots wasn’t just philosophical. Over the years we had numerous requests by policy-bound or very (justifiably) paranoid customers to run their own infrastructure that could be logically separated from ours as much as possible

I dunno. I'll have to see what it turns into. But it surely explains the silence and lack of updates on zt/zt-one

[api]

LF is part of a return to focus: global scale network coordination and decentralization. Making a hardware box was the big defocusing project and took more time (with less results) than this. Lesson learned: avoid hardware if you are not a hardware company in your DNA, and hardware will always take 10-20X longer to ship than to prototype.

So yeah, defocus is a legit criticism but not because of this project.

This was also a two birds with one stone project in that it is also useful on networks with arbitrary connectivity graphs and unreliable connectivity. An enterprise customer of ours funded some of this R&D for that purpose and we combined their objectives with our root server decentralization and network robustness improvement goals. I wish I could talk about the enterprise customer's project. It's quite cool but we're under NDA. All I can say is that it's industrial and involves mobile ad-hoc networks.

There is admittedly quite a lot behind this that is not apparent from the README, and from the comments here its clear that people have trouble understanding this because it is not yet another cryptocurrency and isn't solving exactly the same problems. (There is some overlap but this is different.) We built this to solve a problem nothing else solves, but I figure it will take the appearance of use cases for people to get what that problem is. The people bringing up etcd and consul are closer than the people bringing up coins.

[rootless]

> We built this to solve a problem nothing else solves, but I figure it will take the appearance of use cases for people to get what that problem is.

Beautiful solution. Adds so much to what ZT can provide for borderless networking. First thing I'd implement is distributed DNS for usability. Then probably the config backend for the network layer.