no specific flood fill

or whatever.
but cumulatively?
I mean, if we look at say ipv4.
typically you've got a default gateway on clients.

currently the flood fills know all peers. but not in the future

and then that next hop might go to some bgp router; which has a routing table of various netblocks to pass a packet onto.
xtoaster: *nod* yeah I'd expect it could be broken down a bit more; but just trying to understand the paradigm a bit.

reqiring i2p for torrenting is a lot like requiring no leeching, and in that 1 intance, that makes sense

Goldstein: oh sure.

possiblely there they will use a semi-DHT mechenism to infor peer of each other.

like we said earlier; even peering at internet exchanges has concepts of ratios.
xtoaster: *nod* that could be viable...
with any of these, scaling is a challenge.
not to say it's impossible.
just hard. :)

yep esp in hostile environment

yeah, meshing and scaling is hard enough.
doing it with anonymity as a requirement...
just makes it that much harder.
still seems interesting.
that said,
protocols like bittorrent work much better over things like this, even tor.
than a traditional protocol over a meshed network.
same sorta problem you see with ipv6 tunneled over ipv6.
suddenly if you're doing OSPF or something.
maybe it's only 3 hops of ipv6
but tunneled over 10 hops of ipv4.
and since tcp stream protocols will suffer.
since bittorrent and ilk open many connections.

speaking of ipv6, how is the great ipv6 experiment going?

even if some of them are high latency, or whatnot...
you'll still get overall a pretty even distribution of traffic.
even with anonymity thrown in.

i2p use packet switching instead of circuite switching

so I'm definitely a fan of p2p systems for making anonymity routing networks perform better.
xtoaster: also, a big advantage as tcp has shown us relative to telcom. ;)

:)

iirc my gripes with i2p at the time were java...
and I don't think it had socks support
but that might've been a long time ago.
it looks like that's changed now.

grey its has now

yeah, that's a big plus.
their old stance on it didn't make sense.

why is that so important?

it's very common for an application to support SOCKS.

easy to port i guess.

yeah

obviates the need to alter a program to support your standard, or to wrap it.
that was one of the huge wins for tor long ago.
compared with say something like MUTE.
trying to be everything; and failing at being good at any of them.

but for example, could a torrenting client with socks support use i2p's network even with socks?

I would assume so.
but I haven't touched i2p in so long as that wasn't even configurable.
maybe I'll give it another look now. :)

if you mean to internet, maybe. if you mean in net torrenting , no.
they still use key as destination address :(

yeah, so that's dumb

hmmm.
does i2p have a form of hidden services?

yes

its a network of hidden service :)

like, just creating an anonymous irc or silc server over i2p could be handy.

its more matural actually

every thing hidden service

but the socks support means I can use my preferred irc client say.

xtoaster++

xtoaster: heh, fair enough.

yep. it through i2ptunnel

only if someone outproxies probably

ok, this is interesting.

destination 2 destination

but I gotta go.
ttyl!

ok cya

cu
I mean i2p has like 3 irc servers...
and today also spongue said that he had a very low cpu usage with java on an old machine
it is really getting better and worth taking a look

do bridges with more bandwidth get handed out more?

which leads to the more general question
what is the bridge allocation policy

they are working on better policies

It's occurred to me that they must not have much sense real-time of how much traffic a bridge is getting.

discussion of possibilities for allocating bridges https://blog.torproject.org/blog/november-2009-progress-report

well, your bridge publishes to a bridge authority, so that authority might know everything

cripes this page chews up a few hundred MB in firefox https://www.torproject.org/bridges

big flash video?

ogg i think

oh no, someone changed 0.2.2.7 to 2010-01-?? :(

https://git.torproject.org/checkout/tor/master/doc/spec/proposals/128-bridge-families.txt
hmm
and then theres proposal 125, but nothing on bandwidth that i can see

ok, thx

hmm, is the git server git.torproject.org down?

or: pootle committed revision 21389 (/translation/trunk/projects/website/de): Commit from The Tor Translation Portal by user carolyn. 23 of 23 messages translated (0 fuzzy).
or: pootle committed revision 21390 (/translation/trunk/projects/website/zh_CN): Commit from The Tor Translation Portal by user carolyn. 23 of 23 messages translated (0 fuzzy).

Hello. I'm searching details of current bandwidth evaulation process. Could you help me how can i find it?
I've read A-Tune up for Tor paper but it mentioned current bantwidth evaluation process a little.
It says only: a bandwidth value that is determined by monitoring the peak bandwidth achieved by the router over a period of time.
How long is this period? How is making monitoring to determine bandwidth value for a router?

tor?

toddf: yes, we're working on getting it set up again

hi. There is a 404 on the torproject webside
The first link on http://www.torproject.org/docs/tor-doc-relay.html.de labeled with "Bandbreitenlimitierung, flexible Exitpolicies und Unterstützung für dynamische IP-Adressen." results in a 404.
It seams, the German FAQ is missing completely. If you are at http://www.torproject.org/faq.html.en and click on the German flag you get an empty page with the standard footer.
furthermore the links from https://wiki.torproject.org/noreply/TheOnionRouter/TorFAQ labeld "Answer moved to our new FAQ page" are broken.

why dosn't get my bridge any traffic?

plopix: we don't give out every bridge immediatly
that way, we still have a reserve of bridges in case the censors learn many
copton: will add it to my todo list

Sebastian: ah, ok. thats an explanation.
BTW: I configured a bridge because of the 26C3 presentation. ... just to give you feedback ;-)

plopix: awesome!
plopix: unfortunately, it is much harder to verify that a bridge is working
compared to a relay

could plopix just connect to his bridge himself to see if it works?

yeah Sebastian this came to my mind, cause now I have such a low bandwith usage of tor, that I wasn't sure if it is working, although I guess everything is configured fine...

copton: that is an option, but it doesn't tell you if your bridge is censored
copton: but it is a way to verify that the config is sane.

hi, if I can't connect to tor, the vidalia control panel stops with the status: loading network status

Hello. I'm searching details of current evaulation process for bandwidth value. Could you help me how can i find it?

find ./* -name "*.c" -exec egrep -nH "bandwidth*" {} \; :)

Guys, any tips on why RelayBandwidthRate 2 MB and RelayBandwidthBurst 3 MB don't seem to work with 0.2.1.21? Currently relaying at 12MBs - Even though "we expect to exhaust our quota for this interval around 2010-01-23 13:11:24" on is on the log file, it's not working.

OFFShare: is there a timeout if I'm connected to tor and the connection is not used?

on IRC , can be. On HTTP yes.

gouki: 2MB = 16MBit

OFFShare: k thx

gouki: what exactly isn't working?

Sebastian, d'oh!! s/MB/Mb
Sebastian, don't know how I missed that. I've replaced it with Mb, instead of MB, and it should be working fine now.
Thanks!!

gouki: not sure, I think the options are case-insensitive
so please convert it in megabytes

re

'morning
gouki: tor counts in full units. not fractions of units. there is no 'megabit' in tor-land.

evening armar

Ganneff i want be banned to the server fu*keur! 50 000 minute!

what happens if any bandwidths[i] == 1 (taken from smartlist_choose_by_bandwidth()) as example number of relays is 1000, but only those relays only has Exit and Guard flags let is 50/50 percentage. And we chose with NO_WEIGHTING. I guess total_bw become zero as result. I mean edge cases are presents for any hardware.

who know about console tor-controller?

arm, you mean?

arma: can i switch tor-node from arm ?

eG1k: oh, you want a command-line controller for a client. no, there is no simple one.

thx
https on git.torproject.org not work for download spec to protocol(

google for TorCtl.py
That may/not provide an approximation of what you need.
http://fscked.org/proj/torctl/pydocs/

do you just want a new exit node? you can telnet to the control port and send a NEWNYM

ok)

hello, i have a question if anyone is available

ask and it shall be considered

i just downloaded the vidalia bundle after upgrading to windows 7, and when i used to be on vista, i believe that privoxy was included with the bundle, but now on windows 7, the bundle doesnt seem to include privoxy, or at least windows isnt showing the P tray icon like it did in vista
i downloaded privoxy by itself, and when i try to open it it says 127.0.0.1:8118 is being used by another proxy app, which i assume is vidalia

no, that would likely be polipo
which replaced privoxy in recent bundles

aha!

polipo doesnt have a tray icon afaik
but it seems to be running, if it is using that ip and port

ok, so if i want to route my traffic through a proxy connection, polipo will do it through 127.0.0.1:8118?

yes

polipo will handle http and related dns traffic

you should basically notice no difference compared to privoxy, as long as you're using Torbutton

yeah, i only configured privoxy the once and left it alone after that, but this is good
ok, thanks for your help, thats all i really needed to know

great
i think the install doesnt remove privoxy if they had an older vidalia bundle

not sure, but probably
our upgrade paths for bundles aren't that great
one thing we'll need to solve for Thandy.

well its easy for the portable browser bundle, just erase the whole folder tree

which makes it non-trivial again, since Thandy is running from that folder, too.

hehe upgrading ones elf is tricky
virus-like behaviour

Hello, I want to know if using a broadband connection with tor, for example, for web-browsing, the speed of the broadband connection will likely affect the speed of the web-browsing through tor?

yes, why wouldnt it?

for example, if my broadband ISP offers multiple (speed) tiers of service, will the highest tier (highest speed) be faster for web-browsing than the lowest tier (lowest speed)

yes, it will affect your speed, but you will not see any kind of broadband bandwidth with Tor.

i'm asking b/c does the overhead of the tor network make any increase in the speed of the broadband connection a marginal improvement in speed with the tor network?

Tor is especially if your connection is

no tor will be your bottleneck

erm
Tor is especially *slow* if your connection is very poor
so you will see some huge speed gains, but Tor will still be slow.

your connection would basically have to be dialup speed for it to be your bottleneck

what about 128Kb/s which is typical low-end broadband

> 56.6K

if you want to run a relay however you must have at least 20KB/s (upload+download) to spare
s/must/should/

Goldstein: one would naively think that, but that isn't true for the current Tor implementation. It works dramatically better when you have broadband.

hmmm... i seem to be getting conflicting responses here... so, let me make my example more concrete: say my current speed tier offers {7000 Mbit/sec download, 512 Mbit/sec upload}, but I am considering switching to the lowest speed tier which is {768 Mbit/sec download, 128 Mbit/sec upload}... will this make my tor connection for web-browsing, unbearably slow? Also, for now, I am not running as a relay...

y?

128Mb/s? wow

lol

rip: that will probably not be very noticable.
rip: I'm assuming you're talking about kbit/s ;)