MinimaLT: Minimal-latency Networking Through Better Security (2013) [pdf]

Wasn't there research showing that in channels mixed with tcp traffic , high tcp traffic can effectively block MinimaLT traffic, effectively creating DoS?

This sounds great. A year later, where's the code?

Looks like it will be here: http://www.ethos-os.org/

@JonSolworth (https://twitter.com/jonsolworth) occasionally posts progress updates.

>Whether strangers or anonymous users are allowed is left to the underlying system's authorization policy.

While DoS protection is nice, I would hate to encourage networking that forces client authentication. Imagine if Google/Facebook/etc could effectively enforce a real name policy.

That doesn't matter. You can generate a new Curve25519 identity key per connection.