[tor-talk] comparison of Tor and Kovri in regards to deanonymization attacks

[Eugen Leitl]

I was curious for Monero dev's rationale to pick I2P over Tor, and then even forking I2P as Kovri.

Whatever I've seen online doesn't strike me as particularly convincing.

Is there published research in regards to deanonymization attacks against both Tor 
and I2P, and given the design changes of Kovri, should we expect the attacks to be easier, or harder? 

I realize that the answer would be likely we don't know, which is probably an answer in itself.