@eyedeekay
&kytv
&zzz
+R4SAS
+RN
+altonen
+dr|z3d
+hagen
+lbt
+postman
+radakayot
+snex
+wodencafe
Arch
Danny
DeltaOreo
FreeB
FreefallHeavens_
Irc2PGuest12735
Irc2PGuest28644
Irc2PGuest4450
Irc2PGuest4772
Irc2PGuest59134
Irc2PGuest87589
Onn4l7h
Onn4|7h
Sisyphus
Sleepy
T3s|4_
Teeed
acetone_
aeiou
ardu
b3t4f4c3__
bak83
carried65904
cumlord
death
dr4wd3_
eyedeekay_bnc
mareki2p_
not_bob_afk
onon_1
orignal_
poriori_
profetik1
qend-irc2p
rapidash
shiver_
solidx66
thetia
u5657
uop23ip
w8rabbit
weko_
x74a6
orignal
"For IK: After the 'es' message pattern and before the 's' message pattern, add:"
orignal
Would be better to write full encryption sequence
zzz
orignal, good point
zzz
there's no mixKey() yet
zzz
so you increment n, and for the static key section the chain key remains the same, and n == 1 instead of 0
orignal
so the same key
orignal
and Alice does one x25519 DH and 1 ML decaps
zzz
two DH for IK
zzz
ee, ekem1, se
orignal
then how it works faster than 4?
zzz
it's not faster in total of course, but decaps is faster than DH
orignal
lol. I can make x25519 faster if you let me have 800 bytes key ))
orignal
btw, can you make proposals more clear about speed? because people think that whole thing works faster
orignal
?
zzz
it does say it's slower
dr|z3d
"Note: this will not turbo up your i2p installation."
dr|z3d
*** grins. ***
zzz
MLKEM512_X25519 4x DH + 2x PQ (keygen + enc/dec) = 4.9x DH = 22% slower
zzz
the numbers in there now are simply for our decision making, to make sure we're not making things 100x slower
zzz
but sure, send me a patch with your speed test results, and I'll put them in
dr|z3d
You can forgive the assumption though.. quantum == two locations at once == faster networking because packets arrived at the destination the same time as they were created on the server :)
dr|z3d
</idiot>
zzz
noooo, this is _post_ quantum, you can't even imagine it
dr|z3d
I know, I was being ridiculous for laughs :)
orignal
well people mentioned it was faster
RN
those darn people. they say the darndest things
RN
maybe it is faster if running on qbits... but we are not there yet