~dr|z3d
@RN
@RN_
@StormyCloud
@T3s|4
@T3s|4_
@eyedeekay
@orignal
@postman
@zzz
%Liorar
+FreefallHeavens
+Xeha
+bak83
+cancername
+cumlord
+hk
+poriori
+profetikla
+uop23ip
Arch
DeltaOreo
FreeRider
Irc2PGuest15271
Irc2PGuest28511
Irc2PGuest64530
Meow
Nausicaa
Onn4l7h
Onn4|7h
Over
acetone_
anon4
anu3
boonst
mareki2pb
not_bob_afk
plap
shiver_1
simprelay
solidx66
thetia
u5657
weko_
mesh
I wonder does anybody's router ever participate in much more than a 1000 tunnels?
dr|z3d
of course.
dr|z3d
I've personally seen 14K on a router.
dr|z3d
I'm currently looking into more throughput for participating traffic, mesh. Latest build relaxes the throttles.
mesh
dr|z3d: do you have any idea what drives the participation count? I find it curious that having significantly increased bandwidth limits, system resources like ram, the router struggles to participate for more than a 1000 tunnels
dr|z3d
various things, not least network demand.
dr|z3d
also, router uptime and length of time the router's been known on the network can influence how many tunnels you participate in. the longer you're around, the longer you're up, the more requests you'll get.
mesh
dr|z3d: well I'm not sure I actually want to optimize for throughput instead of latency. I imagine for most people not downloading torrents latency is more important.
dr|z3d
one doesn't have to come at the expense of the other.
mesh
dr|z3d: yeah I've seen that mentioned before. I'll have to dig into the code where router uptime influences these factors.
dr|z3d
uptime == network visibility.
dr|z3d
if you're offline for any length of time, say for more than 1/2 hour, you'll find your participating traffic drops considerably until you've reestablished yourself.
mesh
dr|z3d: What I'd like to actually figure out is the "performance profile" of a i2p floodfill router. Like if I told a stranger to run a floodfill, how much ram should theyu expect to need, cpu, bandwidth, diskspace (neglible).
dr|z3d
start with 512MB and go from there. if they're sharing a good amount of bandwidth, then 1024 isn't unreasonable.
dr|z3d
dual core should be sufficient. disk space isn't an issue.
mesh
dr|z3d: okay. It'll be interesting to see what happens. We'll see where we are in 30 days
mesh
dr|z3d: I'm actually thinking at least 4gb ram is required based on the numbers I've been seeing. My concern with running a floodfill with just 1gb is that your cpus (especially the poor "vCPUs" offered by most hosting providers) wouldn't been able to keep up with the garbage collector
mesh
admittedly haven't actually run a floodfill with 1gb. It might be worth a run. But at the same time it's clear that ZGC is the best gc to use and ZGC does better with larger heaps
albat
hi all :)
parabo
hi guys
parabo
My grand unified field theory is so adopted among researchers who can be trusted, there have been over 300 changes by those who have access
parabo
on a very related note, I have a 800 megaton capable fusion created weapons grade plutonium cache
parabo
I will probably launch it at that rock with an 80% chance of hitting earth causing total liquification of the planet in 2081 that they have been tracking for 3 decades
parabo
I mean, nuclear bombs are good for things other than war or damaging life or the planet