@eyedeekay
&eche|on
&zzz
+FreefallHeavens
+R4SAS
+RN_
+Romster
+T3s|4
+acetone
+cims
+eche|off
+hagen
+mareki2p
+mesh
+nilbog
+nyaa2pguy
+orignal
+postman
+qend-irc2p
+red
+snex
+wodencafe
Arch
Danny
Holmes
Irc2PGuest28384
Irc2PGuest84270
Irc2PGuest92627
NiceBoat_
OfficialCIA_
Onn4l7h
Onn4|7h
Over1
SilentWave
Sleepy
T3s|4_
U1F642
Wikk_0
Zapek
aargh4
ac9f
ahiru
anontor
dr4wd3
duanin2
eyedeekay_
eyedeekay_bnc
ice_juice
leopold
mahlay
makoto
n2
not_bob_afk2
poriori
profetikla
r00tobo
rapidash
test3847473
thetia
uop23ip
urist_
utp
vivid_reader56
x74a6
zelgomer
zzz
eyedeekay, re: branches, was going to delete I2P_Developers/i2p-2.11.0-maven-fix but it does not appear to be merged. Status? Please either merge and delete or just delete.
zzz
hmm datura.network
zzz
Gradual bandwidth allocation over 7 days
zzz
First of all, this feature is there to avoid the sybil attack problem that i2p faces
zzz
PoW-based Resource Allocation (Variable Pricing)
zzz
Nodes limit the bandwidth usage of other nodes to 10kbps by default, until they solve their PoW challenge. (bandwidth unlocking procedures)
zzz
The only way to widen the bandwidth is to solve the RandomX challenge of the given node you want to route the traffic through
snex
i linked that a while ago in a different room
zzz
just saw a tweet about it
zzz
A brand new Darknet that's about to be written in rust, will defeat threats of global passive adversaries, sybil attacks and ddoses, while maintaining bi-directionnal anonymity for both the clients and the hidden services
nyaa2pguy
i had thoughts that it probably wouldn't be hard to implement something like decoy traffic on top of i2p (for bespoke apps)
nyaa2pguy
well, not "too hard"
zzz
sure, the issue is you need a lot to be useful
nyaa2pguy
for one of my many projects i was working on an idea for configurable message/data length randomization to add uncertainty of what people transfer between each other for observers
nyaa2pguy
not quite the same as decoy data to other addresses though
nyaa2pguy
well, it also involved semi fast rotating b32 addresses between peers
snex
why dont you just make a plugin that finds peers with the same plugin via DHT and then sends decoy data
snex
use i2cp or SAM
nyaa2pguy
pretty good idea tbh but the dht values in my poc app are encrypted lmao
nyaa2pguy
probably kinda dumb though: dht key = sha256(sha256(peer id)), dht value = aes256(peer id, peer b32 connection data)
nyaa2pguy
dht key = sha256(sha256(peer id)), dht value = aes256(sha256(peer id) as key, peer b32 connection data)
snex
whats the point of encrypting ephemeral b32s?
nyaa2pguy
stop people from knowing who's online and publishing their b32
nyaa2pguy
also timestamp rounded to day included in the hash calc
snex
huh? the b32 would be thrown away every 10mins or whatever
nyaa2pguy
its like anti spam
nyaa2pguy
my idea is probably a bit excessive though, basically 3x or 4x separate b32s per client (1 for a permanent dht node, 1 or 2 for peer comms (2 when there's a connection handover), 1 extra ephemeral for dht lookups)
snex
you dont need any of that shit
snex
snark opens 1 tunnel and finds DHT peers just fine
snex
basically all you need to do is create snark that restarts itself every 10m and sends fake torrents
snex
just gotta ensure it has low priority so real traffic always goes first
nyaa2pguy
ah yeah that would be better for a torrent client (im not making a torrent client)
nyaa2pguy
well i am making a torrent client but this is for a different dumb idea
snex
the problem youre trying to solve (decoy data) is solved by everyone running a torrent client that just shares fake torrents. you dont need anything more complicated than this
snex
though you probably want to avoid torrent specifically because it has built-in data checking and shit. just do the DHT part and send random data around
snex
stop overcomplicating things
nyaa2pguy
the ideas i had were basically for yet another new collaboration/communication/chat app
nyaa2pguy
with the usual myusername@bla discoverability, on i2p, without people needing to think about what a b32 is
snex
what if you made something we dont already have
nyaa2pguy
what are we missing 👀
snex
we dont have a deadpool
snex
what kind of dark web doesnt even have a deadpool
snex
or just a generic prediction market
nyaa2pguy
that is definitely beyond me lmao
snex
ok how about a lotto
snex
trustless lotto*
snex
there used to be a guy running a lotto on bet.i2p but you had to trust him and then he just disappeared one day
nyaa2pguy
yeah i couldnt be trusted with that either lol
nyaa2pguy
that datura spec is insane, forwarding the traffic 8 times lol
zzz
yeah I skimmed it I have questions
zzz
on twitter, I invited both the datura folks and some other guy doing something over here
nyaa2pguy
a bit puzzled how it solves sybil deanon attacks, unless it's not the "we run all the nodes" kind, but the "we can observe all nodes" kind
snex
it doesnt "solve" it so much as make it very expensive to carry out
snex
or at least more expensive
nyaa2pguy
yeah i guess for tor it's just too cheap to do, all you need is netflow logs for 8000 tor routers. which the intelligence agencies do have.
zzz
the PoW stuff seems to be focused on end-to-end negotiation but our Sybil troubles are network layer
zzz
and they want to ramp up the bw for new routers over a week but who tracks who is new?
snex
each peer does it themselves i guess?
zzz
:shruggie:
nyaa2pguy
can't help but think datura hidden services would basically ddos themselves if they were popular, so i wonder if there's a 'multihoming' capability
zzz
that might interact badly with the PoW
eyedeekay
I had an idea a few weeks ago when things were especially bad that we could do a tunnel build request PoW which only activated when a you were locally congested+certain percentage of your netDb was reporting congestion. The idea was that cost-based defenses would kick in during attacks but when attacks were not happening the network could operate normally
eyedeekay
I wrote the first paragraph of a proposal and set it aside to work on other things. If we're thinking about PoW stuff that possibly applies to us, I could finish writing it up.
zzz
hi
zzz
eyedeekay, re: branches, was going to delete I2P_Developers/i2p-2.11.0-maven-fix but it does not appear to be merged. Status? Please either merge and delete or just delete.
eyedeekay
Ack yeah that's why it wasn't deleted just a sec I'll merge it and delete myself now
zzz
ok thx. looks like StormyCloud deleted his branches on gitea but not GH, please delete everything on the GH side
eyedeekay
Ack will do
zzz
I looked for a gitea setting to prevent new branches but didn't find one
eyedeekay
No I don't think there is I could maybe write a hook or something for it but it would be just as likely to cause problems as solve them
zzz
yup
orignal
no reason for encrypted client LeaseSets
eyedeekay
There's a bootstrap problem if each peer does it themselves though, unless they distribute profiles alongside peers then new nodes only see new nodes, nobody's at least a week old until you're at least a week old
eyedeekay
But with the blockchain stuff they seem to be trying to integrate maybe you could maybe use that to give a node an unforgeable birthday