I'm reading this paper and I have difficulty to understand the purpose of the division in two "legs" of a message path.
I understand the first leg contains the path chosen by the sender and the second leg contains the SURB (Single Use Reply Blocks) that allows to recipient to reply and protect against replay attacks.
The crossoverpoint places the SURB in the first leg but what happens after? How the recipient is join and how the SURB is use by the recipient for to send his reply?
