Kovri has decided to fork the i2p protocol stack . I too have had frustrations with the direction and pase of java i2p and I can understand why they made this descision.
In my personal opinion, I don't think kovri will succeed in their fork of the i2p protocol stack with its current talent pool. After their 2 years of shuffling whitespaces around and not noticing code leaking regions of memory (until stiffing a security researcher that found it while exploit hunting) I don't think they'll produce a minimum viable product for a very very very long time, if ever. This is not meant to poke fun at the people at monero and kovri as I really wanted to see the fork succeed, very much so and I am dissapointed in the net result of the effort.
Parallell to all this kovri tom foolery, since Late April 2018 I have made a fully functioning i2p
"redux" from the ground up.
Very much thankfully I have been hired to work on this full time. The protocol stack is called
LLARP which is an acronym for
Low Latency Anon Routing Protocol.
LLARP is an open standard that anyone can implement in their favorate language. It's a packet based onion routing network utilizing bidirectional onion routing
"paths". It uses bittorrent encoding for message serialization with fast cryptographic primatives
x25519 for key exchanges
blake2 for hash/keyed hash and
xchacha20 for symettric cipher.
"protocol docs" are currently a hot mess in the spirit of being just like i2p (i kid), they can be improved if someone complains loud enough.
The reference implementation is in C++ licensed under ZLIB, a very good permissive license with some good verbiage like
"The origin of this software must not be misrepresented; you must not claim that you wrote the original software." which I find quite valuable.
The current reference implementation is called lokinet which has some additional neato features like cryptoeconomic anti-sybil incentives but I digress. The future kovri is aiming for has already been implemented and I welcome anyone that wants to help build it up.