I feel "BGP in Erlang" would be exactly the kind of thing I could implement well - even if it did feel icky having to implement "MD5 Authentication" in 2016.

The problem with those sorts of projects however, is inertia. The average hobbyist rarely ever uses BGP. Large networks and ISPs aren't going to implement my personal project as a critical component to keeping their entire infrastructure online without a very good reason.

This project looks promising, I'm hoping it doesn't suffer this problem.