Something is bugging me for a while, based on the birthdate paradox the probability of clientID collision is 1% for 36 users and it is 100% in 1000 users for each tunnel (default 2 byte client id)
does it make any problem?
based on the code vaydns treats the clientID as the same users and this will be results in mixing packets in outgoing queue.
A possible solution that I could think of is the Client asks for the free ClientID to use after the initial handshake.
If there would be such a mechanism, maybe we could safely reduce the clientID to a number even less than 16 bits! (not sure if it is practical or not though. But usually people create servers for many users and sharing in telegram channels)
Something is bugging me for a while, based on the birthdate paradox the probability of clientID collision is 1% for 36 users and it is 100% in 1000 users for each tunnel (default 2 byte client id)
does it make any problem?
based on the code vaydns treats the clientID as the same users and this will be results in mixing packets in outgoing queue.
A possible solution that I could think of is the Client asks for the free ClientID to use after the initial handshake.
If there would be such a mechanism, maybe we could safely reduce the clientID to a number even less than 16 bits! (not sure if it is practical or not though. But usually people create servers for many users and sharing in telegram channels)