IRCaBot 2.1.0
GPLv3 © acetone, 2021-2022
#ls2
/2022/03/19
zzz I wrote a simulation, with some interesting results
zzz the measured rtt is stable, it converges quickly
zzz call the denominator for the ack delay calc. D, as in min(rtt/D, 150)
zzz call the "real" rtt RRTT
zzz for large D e.g. 8, measured RTT -> RRTT + RRTT/D
zzz for small D e.g. 2, measured RTT -> RRTT + 2 * RRTT/D
zzz i.e. measured RTT is twice the real RTT for D = 2