Timestamps are in GMT/BST.
[0:03] <Hirvox> no, it's not necessary to download the entire file before the node can share chunks of it
[0:07] <TheSeeker> ok, so if someone requests a 1 MB chunk of a large file, and an upload is started from a node with the chunk in their store... then the first node to recieve this chunk will start sending it to the next node in the chain immediately without verifying the integrity of the data?
[0:10] <TheSeeker> wouldn't that lead to someone being able to easily poison the network by randomly or purposfully altering data as it passes through the node? nobody would be able to tell who it was that was corrupting the data, and they would be able to censor (corrupt) anything that passes through their node...
[0:25] * matahar1 (~matahari@queeg.cs.mu.OZ.AU) has joined #freenet
[0:44] * wern (~wern@h102n1fls20o845.bredband.comhem.se) Quit ("L?mnar")
[1:36] * linagee (~linagee@netblock-66-245-227-226.dslextreme.com) Quit (Read error: 54 (Connection reset by peer))
[1:37] * linagee (~linagee@netblock-66-245-227-226.dslextreme.com) has joined #freenet
[1:57] * Bonkers (bonker@dyn-wien-209-200.dyn.columbia.edu) Quit ("Client exiting")
[2:38] * pupok (~r00t@81-178-106-165.dsl.pipex.com) has joined #freenet
[2:46] * matahar1 (~matahari@queeg.cs.mu.OZ.AU) Quit (Remote closed the connection)
[3:09] * linagee (~linagee@netblock-66-245-227-226.dslextreme.com) Quit (Read error: 110 (Connection timed out))
[3:26] * tav (~tavino@62.206.114.179) has left #freenet
[3:40] * Sandman (~s@CPE-24-164-232-98.new.rr.com) has joined #freenet
[3:40] <Sandman> Anyone alive?
[3:40] * robilad (~topic@mpiat2305.ag2.mpi-sb.mpg.de) Quit (Remote closed the connection)
[3:40] <Sandman> /etc/init.d/freenet start gives "[!!]", which means it err'd out. Where's the logfile? I can't find it
[3:44] <TheSeeker> distrobution? any other information that might be useful?
[4:04] * kmkezlr_ (~kmkezlr@195.25.40.253) Quit ("leaving")
[4:10] * Bonkers (bonker@dyn-wien-209-200.dyn.columbia.edu) has joined #freenet
[4:11] * d-ArkAngel (~Robert@213-131-104-86.onyx.net) has joined #freenet
[4:16] * tyesinclair (~tyesincla@170.252.64.1) has joined #freenet
[4:16] * tyesinclair (~tyesincla@170.252.64.1) has left #freenet
[4:18] * nextgens (~nextgens@ALyon-252-1-47-52.w82-122.abo.wanadoo.fr) has joined #freenet
[5:19] <toad_> hmmm
[5:20] <toad_> 100x25@10, with 70 million simulated requests, is still over 99%
[5:20] <toad_> 200x25@10, with 38.5 million simulated requests, likewise
[5:20] <toad_> 38.8 in fact
[5:20] * Hory (~Miranda@82.76.81.56) has joined #FreeNet
[5:22] <toad_> 400x25@10 looks like going around in small circles though: http://amphibian.dyndns.org/simulations3-400x25-avg25.png
[5:24] <toad_> oh and the simulator seems a lot slower on amd64... probably a JVM problem
[5:24] <toad_> might be to do with the -ext.jar lib, which we use a little in the simulator for some unnecessary conversions
[5:30] * nextgens (~nextgens@ALyon-252-1-47-52.w82-122.abo.wanadoo.fr) Quit (Read error: 113 (No route to host))
[5:39] * vsalento (~vsalento@cs136227.pp.htv.fi) Quit (Read error: 110 (Connection timed out))
[5:46] * spaetz (~spaetz@80-218-144-36.dclient.hispeed.ch) has joined #freenet
[5:49] * Gnutz (~nick@69-171-243-209.clspco.adelphia.net) has joined #freenet
[5:50] * Gnutz (~nick@69-171-243-209.clspco.adelphia.net) has left #freenet
[5:50] * Sandman (~s@CPE-24-164-232-98.new.rr.com) has left #freenet
[6:02] * vsalento (~vsalento@cs136227.pp.htv.fi) has joined #freenet
[6:15] * Sandman (~s@CPE-24-164-232-98.new.rr.com) has joined #freenet
[6:48] <Sandman> Anyone alive?
[6:49] * Bonkers dies
[6:50] * sdogi_ (~java@84-50-20-29-dsl.kjj.estpak.ee) has joined #freenet
[6:51] * tyesinclair (~tyesincla@170.252.64.1) has joined #freenet
[6:51] * tyesinclair (~tyesincla@170.252.64.1) has left #freenet
[6:54] * sdogi (~java@84-50-19-194-dsl.kjj.estpak.ee) Quit (Read error: 104 (Connection reset by peer))
[6:58] * guido^pe (~unknown@dsl-213-023-238-133.arcor-ip.net) has joined #freenet
[7:34] * sdogi (~java@84-50-21-16-dsl.kjj.estpak.ee) has joined #freenet
[7:37] * ike (wangcity@221.232.194.176) has joined #freenet
[7:40] <ike> hi
[7:41] <ike> can anyone give me a ref file? thanks
[7:41] <i2p_iip> <Sonax> Hi ike
[7:41] <ike> hi
[7:41] <i2p_iip> <Sonax> ike: You not happy with the one on the project web page?
[7:42] <ike> what's web page?
[7:42] <ike> i can not use freenet now
[7:42] <i2p_iip> <Sonax> http://www.freenetproject.org/snapshots/seednodes.ref
[7:42] <ike> thanks
[7:43] <i2p_iip> <Sonax> anytime
[7:47] * sdogi_ (~java@84-50-20-29-dsl.kjj.estpak.ee) Quit (Read error: 110 (Connection timed out))
[8:17] * FallingBuzzard (~srademach@207.152.112.129) has joined #freenet
[8:25] * spaetz (~spaetz@80-218-144-36.dclient.hispeed.ch) Quit (Read error: 54 (Connection reset by peer))
[8:25] * spaetz (~spaetz@80-218-144-36.dclient.hispeed.ch) has joined #freenet
[8:28] * hackeron (~hackeron@hackeron.user.gentoo) has joined #freenet
[8:29] <hackeron> I am for some reason unable to access freenode from another machine on the network even though I set: %mainport.bindAddress=eth0 and %mainport.allowedHosts=192.168.0.0/8 -- any ideas?
[8:30] <hackeron> I'm using the stable version.
[8:30] <hackeron> bah, I mean freenet, sorry
[8:31] <hackeron> I try 192.168.0.1:8888 from another machine and it just hangs on "waiting for reply" == but if I do lynx 127.0.0.1:8888 on the server, it works.
[8:37] * sanity (~ian@81-178-106-165.dsl.pipex.com) Quit ()
[8:39] * ike (wangcity@221.232.194.176) Quit (Read error: 60 (Operation timed out))
[8:48] * sanity (~ian@81-178-106-165.dsl.pipex.com) has joined #freenet
[8:53] * FallingBuzzard (~srademach@207.152.112.129) has left #freenet
[8:58] * pittaman (~mistery_b@d51A410CA.kabel.telenet.be) has joined #freenet
[9:00] * ike (wangcity@221.232.191.209) has joined #freenet
[9:10] * Hory (~Miranda@82.76.81.56) Quit ("CyberLore.net - Recommendations on the best games, freeware and websites.")
[9:11] * linagee (~linagee@netblock-66-245-227-51.dslextreme.com) has joined #freenet
[9:12] * sdogi_ (~java@84-50-17-178-dsl.kjj.estpak.ee) has joined #freenet
[9:12] <ike> hi i2p_iip
[9:12] * pittaman (~mistery_b@d51A410CA.kabel.telenet.be) Quit ("webtecheren op windows")
[9:12] <ike> can u download the file : http://www.freenetproject.org/snapshots/seednodes.ref
[9:12] <ike> and send it to me by DCC
[9:13] <ike> i can not direct access the website
[9:15] <i2p_iip> <Sonax> ike: I cant send you the file, but why cant you get it? Is sourceforge down, or are you somewhere where it is censored?
[9:15] * pittaman (~mistery_b@d51A410CA.kabel.telenet.be) has joined #freenet
[9:16] <ike> i can access the sf.net
[9:17] * sanity (~ian@81-178-106-165.dsl.pipex.com) has left #freenet
[9:18] <ike> but
[9:19] <ike> can not access www.freenetproject.org
[9:20] <i2p_iip> <Sonax> ike: Cant help you out with that :(
[9:20] <i2p_iip> <Sonax> ike: Cant help you out with that :(
[9:22] <i2p_iip> <Sonax> Toad, anyone: How about getting some kind of "unofficial" mirrors of the snapshots page up. It is not the first time people report problems like this.
[9:22] <ike> i'm in china
[9:22] <ike> so i can not access the website
[9:24] * sdogi (~java@84-50-21-16-dsl.kjj.estpak.ee) Quit (Read error: 110 (Connection timed out))
[9:25] <i2p_iip> <Sonax> ike: nasty... sorry i cant help you out.
[9:27] * Zer0Her0 (~Z@17.255.240.70) has joined #freenet
[9:28] * Sandman (~s@CPE-24-164-232-98.new.rr.com) Quit (Read error: 110 (Connection timed out))
[9:32] <Zer0Her0> Ok first dumb question of the day what's the difference between entropy and freenet?
[9:38] * Zer0Her0 (~Z@17.255.240.70) Quit (Read error: 104 (Connection reset by peer))
[9:38] * Zer0Her0_ (~Z@17.255.240.70) has joined #freenet
[9:39] * Zer0Her0_ is now known as Zer0Her0
[9:39] <Zer0Her0> hope no one tried to answer my question while my network connection was crashing hehe.
[9:51] <Hirvox> entropy is client-level-compatible with Freenet, but the innards are completely different
[9:53] <hackeron> I constantly get 'RouteNotFound, reason: No route found' and things are painfully slow -- is this normal?
[9:54] <Hirvox> I wish I could say no. I assume that you've tried reseeding, and are running the latest build?
[9:54] <hackeron> running 5100
[9:54] <hackeron> no, but I'll have a search for 'reseeding', thanks
[9:55] <Hirvox> reseeding is simply downloading the latest seednodes.ref, putting it to your freenet directory and restarting the node
[9:56] <hackeron> ah
[9:56] <Zer0Her0> Hirvox: so what does that mean to me? could/should i run both?
[9:57] * Zer0Her0 has been out of the freenet loop for a couple years now.
[9:57] <Hirvox> AFAIK the development of entropy has been ceased
[9:57] <Hirvox> they are not compatible at network level, so if you want to access content on both networks, you'll need to run both freenet and entropy
[9:58] <Hirvox> I haven't looked into entropy closely, so I don't know how secure/insecure it is
[9:59] <hackeron> hmmm
[9:59] <Zer0Her0> i'll just stick with freenet. i just seen mentions of entropy in the past and had never heard more then that so i figured i should ask.
[10:00] <Zer0Her0> also i have 3 machines i wish to run freenet on 2 will be nodes(as they are desktops) the third will be a laptop that will be quite mobil and jumping on from various wireless networks...should i even bother trying to run it on the laptop?
[10:00] <hackeron> from a quick check, entropy is under development with frequent new builds
[10:01] <Hirvox> http://entropy.stop1984.com/en/downloads.html claims that the latest build was in August
[10:02] <hackeron> I'll probably try it as well. Freenode is great, but painfully slow and files seem dead 90% of the time.
[10:02] <Hirvox> I assume you mean Freenet. Freenode is this IRC network
[10:03] <Ash-Fox> I tried Entropy, pretty much never worked at well
[10:03] <hackeron> crap, I keep misusing the name :) -- yes, I mean freenet
[10:03] <hackeron> Ash-Fox: freenet works better then?
[10:03] <Ash-Fox> In my opinion: Yes
[10:03] <Ash-Fox> Others tend to disagree for some reason.
[10:05] <hackeron> well, if going to listings, the images there are just dead -- clicking on files returns file dead, etc.. Not too impressed with freenet at all so far.
[10:05] <Hirvox> have you tried Frost?
[10:05] <Ash-Fox> How long have you been a node?
[10:05] <Ash-Fox> and how long has your node been running?
[10:05] <Ash-Fox> And I don't believe there is such a error as "file dead"... yet
[10:06] <Ash-Fox> Usually you can find the files on freenet if you keep rerequesting
[10:06] <Ash-Fox> hackeron, the thing about freenet, it's meant to serve for anonymity, not for a filesharing application
[10:07] <Ash-Fox> ie: If you want to download a bunch of mp3s etc... go use kazaa or such, it's made for that stuff :P
[10:08] <Hirvox> I haven't used stable builds for a long while, so I don't know how active the stable network is
[10:08] <hackeron> Ash-Fox: only installed it about 2 hours ago
[10:08] <hackeron> Hirvox: frost?
[10:08] <Hirvox> http://jtcfrost.sourceforge.net
[10:09] <Ash-Fox> hackeron, leave the node connected over night, you should see some better performence then
[10:09] <hackeron> not file dead but: "Data Not found "
[10:09] <Hirvox> a client application that uses Freenet. It's basically a message board / filesharing application
[10:09] <hackeron> Hirvox: I'll try it
[10:10] <hackeron> Ash-Fox: no, its not just, I'm kinda just browsing around and all the icons and stuff dont show up
[10:10] <hackeron> Ash-Fox: s/just/that*
[10:10] <Ash-Fox> hackeron, have you waited for the entire page to load?
[10:10] <Ash-Fox> hackeron, what browser are you using?
[10:11] <hackeron> Ash-Fox: yes, I have.. One by one, the icons will appear as dead
[10:11] <hackeron> Ash-Fox: konqueror
[10:11] <Ash-Fox> you might want to tell konq to use more connections, like 100 if need be for each item
[10:12] <hackeron> Ash-Fox: err, how? :p
[10:12] <Ash-Fox> hackeron, might want to ask a kde help channel that :P
[10:12] <hackeron> Ash-Fox: what browser do you use?
[10:13] <Ash-Fox> My hack of firefox
[10:13] <hackeron> Ash-Fox: err, how do I tell firefox to use 100 connections?
[10:13] <Ash-Fox> install the "tweak network settings" exstention
[10:13] <hackeron> hmmm
[10:14] <hackeron> shoudnt all this be handled by freenet though?
[10:14] <Ash-Fox> freenet doesn't handle the browser settings
[10:14] <Ash-Fox> nor how the browser works
[10:14] <hackeron> I suppose it makes sense
[10:15] <hackeron> I'll try this jtcfrost thats designed as a client for freenet
[10:20] <toad_> http://kwiki.ffii.org/Cons041221En
[10:20] <toad_> HAH!!!!!
[10:20] <toad_> Poland too... they came through despite all the pressure
[10:21] * kmkezlr_ (~kmkezlr@195.25.40.253) has joined #freenet
[10:21] <kmkezlr_> hi
[10:22] <toad_> hi
[10:35] * kmkezlr_ (~kmkezlr@195.25.40.253) Quit ("leaving")
[10:35] <hackeron> how do I run a permanent node?
[10:35] <Hirvox> just leave it running
[10:36] <hackeron> ah :)
[10:36] <hackeron> what about actually helping the community? -- will everything I download/view be shared?
[10:36] <Hirvox> yes
[10:37] <hackeron> ok, excellent -- I'll keep it on then.
[10:39] <hackeron> also, it takes 256Mb from ram by default -- so until thats full, all data I look at will be stored there? -- can I set cache to be stored on disk rather than in ram? (dont have much ram)
[10:39] <Hirvox> that's the datastore, and it is stored on disk
[10:40] <Hirvox> but unfortunately Freenet also requires a lot of ram
[10:40] <ike> i download the http://www.freenetproject.org/snapshots/seednodes.ref
[10:40] <ike> but i can not import it into my freenet
[10:40] <hackeron> Hirvox: its taking over 30% ram -- thats unacceptable :(
[10:41] <ike> in normal settings
[10:41] <ike> i click "Import new node ref"
[10:41] <Hirvox> ike:just put it into the freenet directory and restart the node
[10:42] <ike> i use windows
[10:44] <ike> i try now
[10:48] <hackeron> if freenet is anonymous -- doesnt that mean that any transfer goes through the main server?
[10:48] <Hirvox> there is no main servers in Freenet, it's completely decentralized
[10:50] <hackeron> Hirvox: That means if you just run tcpdump while file is downloading -- you can see the IP number of who's sending it and persecute him?
[10:50] <ike> i use build 5100 in windows and the status is 40%
[10:51] <ike> what's the next for ?
[10:51] <ike> i not sure in this time if i'm in freenet
[10:52] <hackeron> Hirvox: I mean say the RIAA goes on freenet, sets a bunch of files to download and logs what IP they are coming from, easy as that?
[10:52] <Hirvox> hackeron:of course you can see the next node's IP, there's no way around that
[10:52] <hackeron> Hirvox: so its not anonymous at all...
[10:52] <hackeron> Hirvox: not anymore than the edonkey network atleast.
[10:53] <Hirvox> hackeron:you're comparing apples and oranges
[10:53] <hackeron> Hirvox: yes, I am -- but I'm comparing if they are both fruits, or is one a vegetable :)
[10:53] <Hirvox> with Freenet, you don't get the whole file, you'll get it in chunks
[10:53] <hackeron> Hirvox: same with edonkey...
[10:53] <Hirvox> and each chunk comes from a different node
[10:53] <hackeron> Hirvox: again, same with edonkey
[10:54] <Hirvox> and the nodes that supply the chunks just relay them from the other nodes in the chain
[10:54] <Hirvox> so the attacker can't figure out which node is the original source
[10:55] <Hirvox> nor does he/she/it see what you're downloading, because node-to-node links are encrypted
[10:55] <hackeron> Hirvox: again.. You queue "britney spears album" and run tcpdump -- you get a nice IP list to persecute :)
[10:56] <ike> Couldn't connect to the network. Are you sure you have configured Freenet correctly? Also make sure that you are connected to the internet.
[10:56] <ike> maybe i can use freenet by proxy?
[10:56] * kevloral (~kevloral@CZ1-RAS-8-u-0179.du.onolab.com) has joined #freenet
[10:57] <kevloral> g'afternoon all
[10:57] <Hirvox> hackeron:as long as you use Internet as a transmission medium, there's no way around the fact that you must know the other nodes' IP addresses
[10:57] <Hirvox> afk, food
[10:57] <hackeron> Hirvox: exactly. So its no more anonymous than any other network that shares chunks rather than files.
[10:58] <hackeron> Hirvox: edonkey is also decentralized, although naturally quite a different purpose. So both are great -- just not anonymous.
[10:58] <i2p_iip> <Sonax> hackeron: But the nodes on the list need not have the file you requested. They could just be forwarding it.
[10:59] <i2p_iip> <Sonax> On edonkey the nodes you connect to have the data you get from them.
[11:00] <ike> i2p_iip:i can not direct access some site like freenet
[11:00] <ike> how can i use it?
[11:00] <i2p_iip> <Sonax> ike: Is your node running? Can you accees http://localhost:8888/ ?
[11:01] <ike> in this page:http://127.0.0.1:8888/servlet/nodeinfo/
[11:01] <hackeron> i2p_iip: hold on, so data is routed through nodes on freenet without nodes actually ever having it?
[11:01] <ike> the freenet tell me it's running
[11:01] <i2p_iip> <Sonax> hackeron: Sometimes they have it (or parts of it), sometimes they just route.
[11:01] <ike> and i can access localhost:8888
[11:01] <hackeron> i2p_iip: oh, thats excellent then. And you cant exactly sue a router :p -- great stuff.
[11:02] <Hirvox> back
[11:02] <ike> but i can not watch the picture in this page
[11:02] <i2p_iip> <Sonax> ike: http://localhost:8888/servlet/nodeinfo/networking/ocm <-- how many open connections?
[11:02] <ike> Couldn't connect to the network. Are you sure you have configured Freenet correctly? Also make sure that you are connected to the internet.
[11:03] <i2p_iip> <Sonax> ike: if you just started up freenet it takes some hours before it is usable.
[11:03] <i2p_iip> <Sonax> have patience.
[11:03] <i2p_iip> <Sonax> have patience.
[11:03] <Hirvox> hackeron:i2p_iip is a relay bot, Sonax is the real nick
[11:03] <ike> Connections open (Inbound/Outbound/Limit) 2 (0/2/200)
[11:03] <ike> Transfers active (Transmitting/Receiving) 0 (0/0)
[11:03] <ike> Data waiting to be transferred None
[11:03] <ike> Total amount of data transferred 612 KiB
[11:05] <ike> it's i connect with freenet now?
[11:05] <i2p_iip> <Sonax> ike: That is not good, but not to bad for a new install. Give it a couple of hours running.
[11:06] <ike> your means is: i not need install again, just waiting for hours?
[11:07] <i2p_iip> <Sonax> ike: Yes, installation is over. You have done that. Now you just wait...
[11:07] <Hirvox> hackeron:whenever content is requested, it's copied to some of the nodes that are relaying it (approximately every fifth), the attackers can't just attack any node that has the offending content, because the act of requesting it distributes it further
[11:08] <Hirvox> due to the same reason, Freenet is also resistant to the slashdot effect. Popular content gets spread around quickly
[11:11] <hackeron> Hirvox: yeah, I get it now, thanks for that.
[11:12] * kevloral is now known as kevlAway
[11:12] <ike> if it tell me it's running it's means it already can connect with freenet?
[11:13] * tyesinclair (~tyesincla@170.252.64.1) has joined #freenet
[11:13] * tyesinclair (~tyesincla@170.252.64.1) has left #freenet
[11:13] <Hirvox> well, you did have open connections, didn't you?
[11:13] <ike> you know i china there is some site can not access direct
[11:13] <ike> you know in china there is some site can not access direct
[11:14] <ike> like www.bbc.co.uk i can not direct access
[11:14] <ike> so, i not sure, the software if already connect with freenet
[11:15] <ike> but the status is 40%
[11:15] <ike> but now the status is just 3%
[11:15] <ike> how about the status?
[11:17] <i2p_iip> <Sonax> ike: Don't think to much of the status. The open connections page is the important one. You want the number of connections to rise. At some point the images at the web-interface will pop up.
[11:22] <ike> ok
[11:22] <ike> Connections open (Inbound/Outbound/Limit) 2 (0/2/200)
[11:22] <ike> Transfers active (Transmitting/Receiving) 2 (1/1)
[11:22] <ike> Data waiting to be transferred None
[11:22] <ike> Total amount of data transferred 1,582 KiB
[11:22] <ike> this kind of connections is too little maybe ?
[11:24] <i2p_iip> <Sonax> ike: Well integrated nodes has ~ 100 open connections, so you should hope for a few more. But it will come if you give it time.
[11:24] * moskau23 (~Miranda@dsl-082-082-234-101.arcor-ip.net) has joined #freenet
[11:25] <ike> ok i will waiting for it better
[11:26] <ike> i want to know if in china, there is so little people use freenet,
[11:26] <ike> maybe the speed in here will more slow
[11:29] <Hirvox> you're lucky to be able to connect at all, I've heard that the chinese government doesn't like Freenet at all
[11:33] <ike> yes
[11:34] <ike> china government is already block a lot of site
[11:35] <ike> but as you know chinese people is very intelligent
[11:35] <ike> :)
[11:35] <i2p_iip> <Sonax> gotta go. Later
[11:35] * sanity (~ian@81-178-106-165.dsl.pipex.com) has joined #freenet
[11:36] <sanity> poland blocked the swpat directive :- http://kwiki.ffii.org/Cons041221En
[11:36] <sanity> my new favourite EU country :-)
[11:37] <i2p_iip> <Michelle> I thought I was your favourite EU Country :(
[11:37] <ike> swpat is what?
[11:38] <Hirvox> software patents
[11:41] * moskawa23 (~Miranda@dsl-082-082-229-183.arcor-ip.net) has joined #freenet
[11:46] * Hory (~Miranda@82.76.81.56) has joined #FreeNet
[11:46] <hackeron> probably wrong place, but anyone uses entropy here?
[11:51] <ike> there is some software like freenet can access site anonymouse and encryption\
[11:52] <ike> what is entropy?
[11:52] <Hirvox> try i2p
[11:53] <ike> ?
[11:54] <ike> www.i2p.net?
[11:54] <Hirvox> yes, that's the one
[11:56] * moskau23 (~Miranda@dsl-082-082-234-101.arcor-ip.net) Quit (Read error: 110 (Connection timed out))
[12:01] * moskawa23 (~Miranda@dsl-082-082-229-183.arcor-ip.net) Quit ("Miranda IM! Smaller, Faster, Easier. http://miranda-im.org")
[12:11] <hackeron> woah, entropy looks nice. Any particular I should avoid using it together with freenet?
[12:14] <TheSeeker> because running two networks that eat up 100% of your computer resources are bound to cause problems for eachother? (guess)
[12:14] <hackeron> well, why would I use freenet over entropy? -- entropy looks cleaner and feels faster. Definitely far easier to get going and friendlier on resources.
[12:15] <TheSeeker> dunno, I haven't used entropy before.
[12:15] <hackeron> give it a try, quite impressed so far.
[12:17] * spaetz (~spaetz@80-218-144-36.dclient.hispeed.ch) Quit ("Verlassend")
[12:18] * Zer0Her0 (~Z@17.255.240.70) Quit ()
[12:20] * d-ArkAngel (~Robert@213-131-104-86.onyx.net) Quit (Read error: 110 (Connection timed out))
[12:23] <Ash-Fox> entropy doesn't give anonymity like freenet does
[12:23] <Ash-Fox> hence, something else
[12:24] <hackeron> Ash-Fox: err, it says it does..
[12:24] <Ash-Fox> hackeron, still doesn't give anonymity like freenet does
[12:24] <hackeron> Ash-Fox: meaning what?
[12:25] <Ash-Fox> meaning it's different?
[12:25] <hackeron> Ash-Fox: different how
[12:26] <Ash-Fox> on entropy, it's possible to trace where a file came from apparently
[12:26] <hackeron> Ash-Fox: apparently :)
[12:30] <hackeron> ah, entropy is compatible with freenet, but not the other way around, interesting.
[12:36] <TheSeeker> not really
[12:36] <TheSeeker> entropy supports FCP client access, but it doesn't interoperate with freenet...
[12:37] <TheSeeker> so you can use soem old version of Frost to upload content into Entropy, but cannot see any of the content that's on Freenet...
[12:37] <sanity> IIRC entropy doesn't have a very scalable search algorithm
[12:37] <sanity> but it is nicely implemented
[12:38] <sanity> its a shame the entropy people did not use freenet's approach
[12:39] * greycat (~wooledg@192.35.79.70) has joined #freenet
[12:39] <hackeron> "This syntax was chosen for ENTROPY to enable application compatibility between ENTROPY and the freenet network"
[12:40] <hackeron> hmm
[12:40] <sanity> hackeron: yeah, but that is only the Freenet Client Protocol, the user-interface to Freenet
[12:40] <sanity> hackeron: its not what matters
[12:40] <hackeron> sanity: it says network though
[12:40] <sanity> hackeron: it is wrong
[12:40] <hackeron> hmm, ok
[12:41] <sanity> hackeron: they speak the same language to clients, but underneath they are very different
[12:41] <sanity> freenet has a good algorithm, but suffers from bloat because it is a very research-orientated project
[12:42] <sanity> entropy has a nice implementation of a much less ambitious routing algorithm
[12:42] <sanity> combining freenet's algorithm with entropy's neat implementation would be ideal
[12:42] <hackeron> well, its easier to setup, cleaner, doesnt eat up ram, doesnt eat up CPU, supports a nice selection of encryption protocols and still has quite impressing routing, no?
[12:44] * Zer0Her0 (~Z@17.255.240.70) has joined #freenet
[12:44] <sanity> hackeron: well, i agree it is cleaner and less bloated, but it isn't scalable, and is much less ambitious than freenet
[12:45] <hackeron> sanity: less scalable? how so?
[12:45] <sanity> hackeron; let me refresh my memory...
[12:46] * Zer0Her0 (~Z@17.255.240.70) Quit ()
[12:48] * Zer0Her0 (~Z@17.255.240.70) has joined #freenet
[12:51] <ike> hackeron: can you tell me the entropy's website?
[12:53] <sanity> hackeron: freenet forms a small world network with essentially logarithmic scalability. entropy's scalability characteristics are rather vague (but probably not good)
[12:53] <hackeron> ike: http://entropy.stop1984.com/en/home.html
[12:55] <sanity> hackeron: the problem with entropy is that they copied some good ideas from freenet, and did a really nice implementation of them, but they didn't copy the really important good ideas :-)
[12:56] <sanity> hackeron: if they had, i would probably be an entropy user right now (leaving freenet for research)
[12:56] <hackeron> sanity: I still didnt say which good ideas they are :)
[12:56] <hackeron> sanity: err, you still didnt say I mean
[12:57] * Zer0Her0 (~Z@17.255.240.70) Quit ()
[12:57] <TheSeeker> freenet "would" form a small world network if it never had to send any data through it that wasn't for routing purposes... but since anonymity is the #1 goal here, you can't have the person with the data giving it to the person who requested it directly, and thus stuffs up the network with 'trailers' and mucks up routing...
[12:58] <sanity> TheSeeker: that makes no sense
[12:58] <TheSeeker> which causes people to open more connections than they can handle in order to find a non-backed off node *now* making them suceptable to getting a lot more requests than they can handle later when nodes aren't backed off, causing them to back off themselves, ensuring that the cycle continues...
[12:58] <sanity> TheSeeker: more unsubstantiated theories presented as fact?
[12:59] <hackeron> TheSeeker: do you use entropy?
[12:59] <TheSeeker> no
[12:59] <sanity> TheSeeker: i'm not saying you aren't right, you might be, but please don't state your theories as fact without evidence
[13:00] <sanity> TheSeeker: it isn't helpful
[13:00] <TheSeeker> #1 cause of overload: exceeding my upstream bandwidth limit. I can sto pthis by having fewer conenctions, but then everythign will RNF, because I'm the only one lowering my number of connections in order to provide a better service to the network...
[13:00] <i2p_iip> <Michelle> sanity, you are beautiful
[13:01] <TheSeeker> the onyl way as an end user I can avoid RNFs at this point is to connect to more nodes, which eventually makes me go into an overload state myself.
[13:04] <TheSeeker> when I can never choose the best route for any request because most of the nodes I'm connected to are backed off, it means that others are unable to get a feel for what I am capable of routing to best, since I'm basically 'random routing' every request I get (I know I'll choose the 'best' node that I percieve is specialized closes to that key, but the odds of one being anywhere nearby are pertty random for a given key)
[13:05] <sanity> michelle: i know, my mother told me so
[13:07] <sanity> TheSeeker: well, cause of overload is having too many incoming requests, which you then reduce using rate limiting
[13:07] <toad_> <hackeron> I am for some reason unable to access freenode from another machine on the network even though I set: %mainport.bindAddress=eth0 and %mainport.allowedHosts=192.168.0.0/8 -- any ideas? -- remove the % ! :)
[13:08] <toad_> uh, somebody email ike a zip?
[13:08] <toad_> or even make a distro servlet...
[13:09] <hackeron> TheSeeker: yes, figured out already :)
[13:09] <TheSeeker> you also get fewer infoming requests when you have fewer connections, and when every connection requests something of you, you can serve them all with a minimal performance metric (say, 1 KB/s)
[13:09] <hackeron> err, thats to toad_ TheSeeker
[13:09] <TheSeeker> I figured that...
[13:09] <hackeron> toad_: have you tried entropy?
[13:09] * root (~r00t@ool-18be3a49.dyn.optonline.net) Quit (Read error: 110 (Connection timed out))
[13:10] * Zer0Her0 (~Z@17.255.240.70) has joined #freenet
[13:13] <toad_> ike here?
[13:14] <toad_> entropy has home grown crypto algorithms (which suck)
[13:14] <toad_> also it isn't really client compatible; the SSKs don't work
[13:14] <ike> yes
[13:15] <ike> toad: your means the entropy is not really client compatible ?
[13:16] <toad_> <TheSeeker> #1 cause of overload: ?exceeding my upstream bandwidth limit. ?I can sto pthis by having fewer conenctions, but then everythign will RNF, because I'm the only one lowering my number of connections in order to provide a better service to the network...
[13:16] <toad_> you are SUPPOSED to use ~ 100% of your limit!
[13:16] <toad_> <sanity> TheSeeker: well, cause of overload is having too many incoming requests, which you then reduce using rate limiting
[13:16] <toad_> not really, rate limiting lets you move load around...
[13:17] <toad_> hackeron: no
[13:17] <ike> if i want to use freenet for access some website -- be blocked
[13:17] <ike> how can i do?
[13:17] <toad_> ike: well, kinda. one fundamental thing we use is SSKs.. entropy lets you have SSKs but they're not secure crytographic identities like they are on freenet
[13:18] <greycat> freenet is not a web proxy
[13:18] <toad_> ike: you can't, sorry, freenet is not a proxy
[13:18] <toad_> ike: what seednodes are you using? 2 conns is pretty bad
[13:18] <ike> just for p2p ?
[13:18] <toad_> how many connections do you have now?
[13:18] <greycat> that would be a very useful service, if someone were to do it, though.
[13:18] <toad_> ike: well kinda p2p
[13:18] <toad_> web sites, files, message boards, more files
[13:18] <toad_> the web interface is good for the first two
[13:18] <greycat> ike: freenet is used to publish anonymously, or to read anonymously
[13:18] <toad_> frost (jtcfrost.sf.net) is good for the latter two
[13:19] <ike> Connections open (Inbound/Outbound/Limit) 2 (0/2/200)
[13:19] <ike> Transfers active (Transmitting/Receiving) 0 (0/0)
[13:19] <ike> Data waiting to be transferred None
[13:19] <ike> Total amount of data transferred 4,089 KiB
[13:19] <greycat> but it can only retrieve data that's been explicitly inserted into the freenet network
[13:19] <toad_> you can establish a secure pseudonym
[13:19] <toad_> that's what entropy doesn't do iirc
[13:19] <toad_> at least not with ssks
[13:19] <toad_> ike: what seednodes list did you use?
[13:19] <toad_> ike: and what's your bandwidth?
[13:19] <greycat> and are you firewalled?
[13:19] <toad_> oh, and you're accessing this channel through a proxy, i hope?
[13:20] <toad_> yeah, are you behind any sort of router?
[13:20] <toad_> apart from the Great Wall?
[13:20] <toad_> i.e. are you able to receive incoming connections?
[13:20] <ike> i not know
[13:21] <ike> i use adsl 1M
[13:21] <greycat> I guess asking him for his computer's IP address would be a bit problematic...
[13:21] <toad_> how did you start up your node?
[13:21] <toad_> please don't tell us your ip :)
[13:21] <ike> can download 120k and up 50k
[13:21] * nextime (~nextime@ns0.nexlab.net) Quit ("nextime has no reason")
[13:21] <toad_> ike: do you have a DSL router, or is it just a DSL modem?
[13:21] <greycat> damn, that's better than mine!
[13:22] <ike> dsl modem
[13:22] <toad_> ike: we know that some people have gotten it to work over there
[13:22] <toad_> ike: okay, so it should be possible to receive connections
[13:22] <TheSeeker> toad: I'm backed off and not accepting any conenctions or requests (often) it's because I'm already exceeding my bandwidth limit. if every node only conencted to enough nodes that each one could have an active request, then all a node woudl ahve to do to "behave" is to not send a request to a node already sending them data.
[13:22] <ike> in china?
[13:22] <toad_> how did you download it? somebody sent you an email?
[13:22] * nextime (~nextime@ns0.nexlab.net) has joined #freenet
[13:22] <toad_> TheSeeker: not accepting connections because of bandwidth? that's wierd
[13:22] <TheSeeker> ... I'm still trying to write a sim that will show that this will still aoow routing to work, but I have to go on the concept of "bandwidth first"
[13:23] <ike> download freenet software ?
[13:23] <toad_> TheSeeker: I haven't proved that routing works in simulation yet :)
[13:23] <ike> i use another way for download it
[13:23] <toad_> ike: yes, did you get it from the website?
[13:23] <toad_> or was it a custom one from somebody here?
[13:23] <ike> yes
[13:23] <TheSeeker> well, that's what it says when I'm backed off... it could be lying ... [Rejecting incoming connections and requests]
[13:23] <toad_> yes you got it from the website?
[13:23] <ike> i download it by myself
[13:23] <toad_> okay
[13:23] <ike> but use some way
[13:23] <toad_> hmmm
[13:23] <toad_> you ought to get more connections than that...
[13:24] <toad_> what build number does it show up?
[13:24] <toad_> on the Web Interface page?
[13:24] <ike> 5100
[13:24] <ike> i download all things from the office website
[13:25] <ike> the install program is bad!
[13:25] <ike> when install it's must download some part from net
[13:25] <toad_> if you get all the bits together in a directory, it will just install them
[13:25] <toad_> without having to download them
[13:25] <ike> but if let the install program direct download, it's can not work !
[13:26] <ike> so,i had to downlaod those part by hand
[13:26] <toad_> you need freenet.jar (named freenet-latest.jar), freenet-ext.jar, seednodes.ref...
[13:26] <ike> yes
[13:26] <ike> i get all thing
[13:26] <ike> now i let it work
[13:26] <toad_> how big is your seednodes.ref in bytes?
[13:26] <ike> but i not know how to get files
[13:26] <ike> about 5m
[13:27] <toad_> well with only 2 connections it will not work well
[13:27] <toad_> and you won't have much anonymity
[13:27] <toad_> try reseeding...
[13:27] <toad_> stop the node
[13:27] <toad_> and replace the seednodes.ref file
[13:27] <toad_> and then start it again
[13:27] <toad_> you still have only 2 connections, right?
[13:28] <ike> yes
[13:28] <ike> but how to reseeding?
[13:28] <toad_> it would really help us to know if those conns are within china...
[13:28] <toad_> but that's hard to know without using whois...
[13:28] <toad_> unless they have domain names...
[13:28] <ike> your menas want to know my ip?
[13:28] <toad_> ike: stop freenet, remove seednodes.ref, restart it
[13:28] <toad_> ike: no
[13:28] <toad_> ike: we don't want to know
[13:29] <toad_> if you can figure out whether the two conns are in china, that'd be useful to know
[13:29] <toad_> but please reseed
[13:29] <leexx-zzz> lo
[13:29] <ike> but if i remove the seed
[13:29] * leexx-zzz is now known as leexx
[13:29] <toad_> ike: this forum may be monitored, and it IS logged, so we don't want to know your IP. Logs can go missing occasionally... but if somebody is watching, you're stuffed
[13:29] <ike> i can not connect with freenet?
[13:30] <toad_> ike: remove it and then get it again and put it there
[13:30] <toad_> this forces freenet to reread it
[13:30] <toad_> once it has read it it no longer needs it
[13:30] <leexx> ike run UpdateSnapshot.exe
[13:30] <toad_> but reseeding helps it to connect to more nodes
[13:30] <toad_> leexx: he can't download directly
[13:30] <leexx> that downloaded the seed nodes for me
[13:30] <leexx> why ?
[13:30] <leexx> ok i read up
[13:31] <ike> i can not use updatesnapshot.exe bcz i can not download directly
[13:31] <toad_> right
[13:31] <leexx> does your dcc work
[13:31] <toad_> you have an up to date version of the jar
[13:31] <toad_> you need to reseed
[13:31] <ike> the updatesnapshot.exe can use proxy ?
[13:32] <toad_> at least that's what I'd do
[13:32] <ike> yes , i can use dcc
[13:32] <toad_> no idea
[13:32] <leexx> dcc him nodes file
[13:32] <toad_> ike: hmm, so you probably CAN receive connections
[13:32] <leexx> hoq then did you get the freenet then??
[13:32] <leexx> how
[13:33] <greycat> (and more importantly, *where*)
[13:33] <ike> i geting now
[13:33] <ike> but i not understand
[13:33] <leexx> if you cant directly download files then how did you download the files (ok)
[13:33] <leexx> if you get me
[13:33] <toad_> leexx: he indirectly downloaded it
[13:33] <ike> i already download a this kind of file in the website
[13:34] <leexx> does freenet support proxys
[13:34] <leexx> even
[13:34] <toad_> the installer? i don't know
[13:34] <toad_> i don't think so
[13:34] <leexx> does freenet
[13:35] <toad_> leexx: hmm?
[13:35] <toad_> leexx: the only reason freenet would need to use a proxy would be if you were behind an application-proxy-only firewall
[13:35] <toad_> we don't support that at present, it'd require tunneling and all sorts
[13:36] <toad_> and it's comparatively rare
[13:36] <hackeron> toad_: entropy uses SHA-1 -- how does it suck?
[13:36] <greycat> that's a hashing algorithm, not an encryption
[13:36] <greycat> what else does it use?
[13:36] <toad_> hackeron: I was told it used home grown symmetric ciphers
[13:37] <hackeron> toad_: you were lied to :p
[13:37] <greycat> hackeron: I'm sure it uses something other than a hash.
[13:37] <hackeron> greycat: well, its covered here: http://entropy.stop1984.com/en/entropy.html
[13:37] <greycat> hashes, while terrifically useful for many things, aren't that good at emulating public key cryptography for inter-node data transfers
[13:38] <toad_> hackeron: okay, but does it have proper public key crypto for SSKs?
[13:38] <toad_> I was also told that their SSKs are not the same as ours
[13:38] <hackeron> toad_: have a look at the link
[13:38] <hackeron> toad_: they might not be the same, but doesnt mean they suck :)
[13:38] <greycat> Each fragment of a file is then hashed using two algorithms: EK5
[13:38] <greycat> (which is a modified MD5 hash for Entropy's encryption key purposes)
[13:39] <toad_> I was told their SSKs do not provide a cryptonym
[13:39] <greycat> and SHA1 after encryption.
[13:39] <hackeron> greycat: yup. Anything wrong with that?
[13:39] <greycat> yeah, I don't trust anything that's "modified for Entropy"
[13:40] <hackeron> heh
[13:40] <greycat> where's the cryptoanalysis for that new algorithm?
[13:40] <toad_> well i don't trust md5
[13:40] <greycat> toad_: good point. :)
[13:40] <hackeron> heh, fair enough
[13:40] <hackeron> ok, I think I've had enough of this entropy -- going back to freenet :)
[13:40] <leexx> is it even been updated any more
[13:41] <greycat> the paper then goes on to talk about an algorithm called "Lili2", with references, but I don't quite understand it at that point. (I'm no cryptanalyst.)
[13:42] <hackeron> do I need to open any ports except for 24651 for freenet?
[13:42] <toad_> you have to open the FNP port
[13:42] <toad_> that's it
[13:42] <greycat> and all outgoing tcp ports must be allowed
[13:43] <leexx> yes
[13:43] <hackeron> greycat: yeah, those are allowed.
[13:43] <leexx> FNP to all
[13:43] <leexx> all to FNP
[13:44] <leexx> incomming rule out going rules
[13:44] <leexx> never used out going limit rule before
[13:45] <hackeron> anyway to make freenet easier on my resources? (ram/cpu)
[13:45] <leexx> as 99% of soft does not give you the option to block out going (in HW)
[13:45] <leexx> waited more than 120000ms in NIOIS.read() tcp/connection: 3181>213.114.107.43:18470,freenet.transport.tcpConnection@1dbe530:freenet.support.io.NIOInputStream@1a3eaa8- closing
[13:45] <leexx> java.lang.Exception: debug
[13:45] <leexx> ffs
[13:45] <greycat> leexx: I don't know what "soft" you're using, but it obviously sucks.
[13:46] <greycat> any competent firewall or router allows you to block in either direction
[13:46] <leexx> most soft firewalls do not limit out going
[13:46] <greycat> wrong.
[13:46] <greycat> unless you're a Windows User or something.
[13:46] <leexx> i mean in an port range way
[13:46] <greycat> IPF, PF, ipchains, iptables all have the ability to block outgoing traffic
[13:47] <leexx> if i sayed to emule only use port 4662 for in and out emule whould not work
[13:47] <leexx> the incomming port is not the out going port
[13:47] <greycat> you're a windows user. I can tell. So it's no use even trying .... :(
[13:47] <leexx> for emule
[13:48] <hackeron> leexx: xmule you mean?
[13:48] <leexx> emule
[13:48] <hackeron> leexx: or do you actually run the thing in wine?
[13:48] <toad_> greycat: PF is not a firewall in itself, it's a way of making firewalls; you have to configure it
[13:48] <leexx> my router is linux
[13:48] <leexx> got linux running in am vmware box
[13:49] <greycat> your router is a windows box that runs vmware that runs linux?
[13:49] <leexx> no
[13:49] <leexx> i got like 7 computers
[13:49] <leexx> 1 router
[13:49] <leexx> 1 server
[13:49] <leexx> 5 norm pcs
[13:49] <hackeron> is there a quick start guide to configuring and using freenet?
[13:50] <greycat> hackeron: lots of them, probably all out of date.
[13:50] <hackeron> greycat: heh
[13:50] <hackeron> greycat: should I use the stable or unstable freenet?
[13:50] <toad_> hackeron: stable
[13:50] <hackeron> well, isnt stable ancient?
[13:51] <leexx> for the most part its moderfilying the freenet.ini to tell it your IP and port (has that been implemated now so it can find its own IP)
[13:51] <greycat> greycat's nutshell guide: make a 'freenet' user, give it a home dir on a disk with lots of space, download freenet.jar, download freenet-ext.jar, download seednodes.ref, install a java runtime environment, "su" to freenet, run "java -jar freenet.jar --config", then run "nohup java -jar freenet.jar &" and pray.
[13:51] <leexx> lol
[13:51] <hackeron> leexx: I dont have no freenet.ini...
[13:51] <greycat> it will be freenet.conf on unix.
[13:51] <leexx> yep
[13:51] <leexx> was thinking of running it on me router
[13:51] <hackeron> greycat: well, I've done that :) -- anything else?
[13:52] <greycat> hackeron: point a web browser at http://127.0.0.1:8888/
[13:52] <greycat> and tail -f freenet.log in ~freenet if you wish
[13:52] <hackeron> greycat: yeah, done that.. but it sucks up all my ram and CPU and hangs from time to time
[13:52] <hackeron> greycat: done that as well :p
[13:52] <greycat> how much RAM and CPU?
[13:53] <hackeron> greycat: 1ghz, 512mb
[13:53] <greycat> that doesn't sound right.
[13:53] <greycat> edit the freenet.conf and lower something like maxThreads
[13:53] <hackeron> greycat: hmm, ok. Will I need to restart freenet after?
[13:53] <greycat> yes
[13:53] <hackeron> ok, lets see..
[13:53] <greycat> (says greycat who hasn't run a node in about 2 weeks now)
[13:54] <hackeron> what are all these?
[13:54] <hackeron> distrib freenet.log freenet.stderr.log lock.lck lsnodes_b ngrt_global_b ratedata_a rtnodes_a rtprops_a seednodes.ref store
[13:54] <hackeron> freenet-downloads freenet.pid freenet.stdout.log lsnodes_a ngrt_global_a node ratedata_b rtnodes_b rtprops_b stats
[13:54] <greycat> "store" is the data store, where all the fun stuff lives
[13:54] <greycat> rt* is the routing table
[13:54] <greycat> the rest is Black Magic
[13:54] <hackeron> heh
[13:55] <hackeron> also, why or why is % comment?
[13:55] <hackeron> why oh why*
[13:55] <greycat> historical reasons.
[13:56] <toad_> so that it can indicate settings that can be updated by tools
[13:56] <hackeron> hmm
[13:56] <hackeron> where's that used historically?
[13:56] <toad_> freenet.config.Setup --silent --somethingelse
[13:56] <toad_> i think somethingelse is update
[13:56] * mikeeusa (~r00t@ool-18be3a49.dyn.optonline.net) has joined #freenet
[13:56] <toad_> that will update %'ed out settings
[13:57] <mikeeusa> HI
[13:57] * jay (~jcl@198.38.10.200) has joined #freenet
[13:57] <leexx> lo
[13:59] <hackeron> startup is very slow :(
[13:59] <greycat> with a big data store, it can be.
[14:00] <hackeron> 2gb is big? -- its still empty though.
[14:00] <greycat> no...
[14:00] <hackeron> still starting...
[14:01] <hackeron> still starting........
[14:01] <hackeron> common already..
[14:01] <hackeron> jumped to 30% cpu usage -- again at 99%
[14:01] <hackeron> started..
[14:02] <hackeron> is this normal?
[14:02] <mikeeusa> yes
[14:02] <mikeeusa> freenet is ummm...
[14:02] <mikeeusa> well lets say
[14:02] <hackeron> now CPU usage is betweeen 10% and 80%
[14:02] <mikeeusa> there's alot of code
[14:02] <mikeeusa> and ...
[14:02] <mikeeusa> well
[14:02] <mikeeusa> it's like a jungle
[14:02] <greycat> it doesn't take a lot of code to eat CPU...
[14:03] <greycat> it should stabilize at a lower level of CPU usage after the initial flurry of activity opening connections
[14:03] <hackeron> fork while fork? :p
[14:03] <hackeron> yes, it did, but has CPU usage spikes
[14:07] <toad_> hmmm
[14:07] <toad_> results for 400x25 with connections on success so far look good
[14:07] <toad_> but i need longer term output...
[14:09] * Hirvox (~hirvox@cs181124095.pp.htv.fi) Quit (Read error: 104 (Connection reset by peer))
[14:10] <hackeron> freenet has resume algorithms like in rsync for example?
[14:10] <toad_> http://amphibian.dyndns.org/simulations3-400x25-avg25.png
[14:10] <greycat> not within a single key, but large files are split into multiple keys
[14:11] <hackeron> greycat: well, say a 5Mb file, how many keys would it be split into?
[14:11] <toad_> a single key is atomic
[14:11] <toad_> 5M would be split up into probably 256kB files
[14:11] <greycat> hackeron: keys can be 1 MB or less...
[14:11] <toad_> and then there's the 50% redundancy
[14:11] <greycat> that's assuming an FEC splitfile :)
[14:11] <hackeron> hmm, I see.
[14:11] <hackeron> WHat about compression? -- freenet uses compression?
[14:12] <greycat> no
[14:12] <hackeron> hmm, ok
[14:13] <greycat> most 5 MB files are already compressed, anyway.
[14:13] <toad_> freenet uses compression for websites
[14:13] <toad_> "ZIP manifests"
[14:13] <toad_> which is where it is needed most
[14:13] <greycat> is that transparent, or does the client application have to do it?
[14:14] <toad_> client, unfortunately
[14:14] <toad_> it's transparent for fetching
[14:14] <toad_> but not for putting
[14:17] <hackeron> hmm, I see. Nice, nice.
[14:17] <toad_> anyone got some seednodes?
[14:18] <hackeron> any tips how to speed things up? -- I heard making browser use multiple connections? -- anything freenet configuration related I can change?
[14:20] <toad_> hackeron: it's in the README :)
[14:21] <toad_> which somebody decided not to ship with the windows version since they'll never read it anyway ;)
[14:21] <toad_> hackeron: you have a functional node? how many connections?
[14:21] <TheSeeker> it's nickserv
[14:21] <hackeron> toad_: how many active or max?
[14:21] <TheSeeker> oops, wrong window
[14:21] <toad_> active
[14:22] <hackeron> toad_: it says Connections open (Inbound/Outbound/Limit)
[14:22] <hackeron> 12 (0/12/350)
[14:22] <toad_> http://amphibian.dyndns.org/simulations3-400x25-avg25.png
[14:22] <toad_> wooh!
[14:23] <toad_> hackeron: why no incoming? you firewalled?
[14:23] <TheSeeker> toad: I have a 400x25@25 that's been running on my laptop all night...
[14:23] <toad_> @25?!
[14:23] <hackeron> toad_: I have a firewall, but I opened 24651
[14:23] <toad_> it's that bad?
[14:24] <TheSeeker> up to 58 million requests now.
[14:24] <toad_> hackeron: what sort of firewall?
[14:24] <toad_> does it do NAT?
[14:24] <hackeron> toad_: iptables
[14:24] <toad_> if so you need to forward the port
[14:24] <hackeron> toad_: yes
[14:24] <toad_> okay
[14:24] <toad_> you need to accept conns on that port, yes
[14:24] <hackeron> toad_: no, I'm on the router
[14:24] <toad_> what's your ip? can i test it?
[14:24] <toad_> http://amphibian.dyndns.org/simulations3-400x25-avg25.png
[14:24] <toad_> that looks distinctly promising!
[14:24] <toad_> red i mean
[14:25] <toad_> obviously green sucks :)
[14:25] <toad_> because the no-connection-churn thing doesn't scale
[14:25] <hackeron> toad_: 81.86.159.146
[14:26] <toad_> trying...
[14:26] <TheSeeker> est-doConns=true ... that the only change?
[14:26] <toad_> TheSeeker: yup
[14:26] <hackeron> toad_: its restarting, give it a few minutes to start accepting requests :p
[14:26] <toad_> doConns means do connection churn
[14:26] <TheSeeker> ah
[14:26] <toad_> i.e. add a connection if you can, i.e. if your nodes are experienced enough to dump one, after a successful request, to the datasource
[14:27] <TheSeeker> my 400x25@25 is hovering around 87% at the moment.
[14:27] <hackeron> toad_: should work now, try
[14:27] <toad_> trying...
[14:27] <toad_> TheSeeker: really? I thought htl didn't make that much difference?
[14:27] <hackeron> my shorewall rule is: ACCEPT net fw tcp 24651 # freenet
[14:28] <toad_> anyway 10 is enough for 100 and 200
[14:28] <toad_> so it's clearly not scaling!
[14:28] <hackeron> listenPort=24651
[14:28] <toad_> whereas with connections, it can scale
[14:28] <TheSeeker> want a copy of that log file? I think I'm going to restart that with est-doConns=true
[14:28] <toad_> the without-connections stuff shows us that it can learn the form of a complex network
[14:29] <hackeron> toad_: I do see firewall doing its magic like: Dec 21 19:34:24 server Shorewall:net2all:DROP:IN=ppp0 OUT= MAC= SRC=65.187.193.170 DST=81.86.159.146 LEN=60 TOS=0x00 PREC=0x00 TTL=46 ID=53092 DF PROTO=TCP SPT=45141 DPT=24651 WINDOW=5840 RES=0x00 SYN URGP=0
[14:30] <toad_> hackeron: so it's dropping connections?
[14:30] <hackeron> toad_: looks like it doesnt it :p
[14:30] <toad_> :(
[14:30] <hackeron> wait a sec.. did I forget to restart it?
[14:30] <toad_> badly configured firewall. refer to system administrator. :)
[14:30] <toad_> probably! ;)
[14:31] <hackeron> restarted it and I see: Dec 21 19:36:10 server Shorewall:net2all:DROP:IN=ppp0 OUT= MAC= SRC=80.28.116.7 DST=81.86.159.146 LEN=60 TOS=0x00 PREC=0x00 TTL=51 ID=12221 DF PROTO=TCPSPT=2344 DPT=16642 WINDOW=5840 RES=0x00 SYN URGP=0
[14:32] <hackeron> Connections open (Inbound/Outbound/Limit)
[14:32] <hackeron> 10 (2/8/150)
[14:32] <hackeron> hmm, progress :)
[14:33] * TheSeeker runs 400x10@10 with connection churn just for fun
[14:33] <toad_> TheSeeker: umm, i'm not sure i committed that...
[14:34] <toad_> hackeron: okay it works now
[14:34] <TheSeeker> well, it didn't balk when I started it...
[14:34] <hackeron> server freenet # echo $(cat /var/log/syslog/kern.log | grep -o DPT=[0-9]* | sed 's/DPT=//' | sort | uniq)
[14:34] <hackeron> 1027 16642 4662 4666
[14:34] <hackeron> hmm
[14:34] <toad_> looking very promising... http://amphibian.dyndns.org/simulations3-400x25-avg25.png
[14:34] <TheSeeker> though you're right, it doesn't seem to be there in the CVS
[14:35] <toad_> once i can show this works, i'll start adding in all the stuff that we thought would improve matters e.g. pcaching :)
[14:36] <TheSeeker> does pcaching just have a random chance of kicking out the LRU key from the datastore and replacing it with one passing through the node? or is it more involved than that (chance to cache based on self-percieved specialization?)
[14:36] <TheSeeker> well, not "random" chance, but a % chance
[14:36] <toad_> chance is based on hops since reset
[14:36] <toad_> which appears to help in simulation
[14:38] <TheSeeker> more hops since reset = greater chance I take it?
[14:38] <toad_> p(cache) = 0.8^hopsSinceReset
[14:38] <toad_> so 1.0 for the zeroth hop
[14:38] <toad_> fairly low for the 10th hop
[14:38] <toad_> 10% chance of caching on the 10th hop
[14:40] <toad_> i'll be VERY interested in what this does at 10M reqs...
[14:40] <TheSeeker> oh, so it has less chance of caching the further it gets from reset? that seems counter-productive to distrobution to me... but perhaps helps specialization stay localized?
[14:40] <toad_> no
[14:40] <toad_> yes
[14:40] <toad_> it helps the specialization stay close to the end of the request
[14:41] <toad_> if it's at the other end, then you end up with severe limits on capacity
[14:41] <toad_> and routing has less to do
[14:41] <toad_> and is therefore less effective
[14:42] <TheSeeker> hmm
[14:42] <toad_> in the unlikely event that this all works out... how do i convince myself that it's not all a mirage caused by a bug i haven't found yet?
[14:43] <TheSeeker> get a lot of other people that know java to pick apart your code and look for errors?
[14:49] * kevlAway is now known as kevloral
[14:49] <kevloral> g'evening all
[14:50] <toad_> hi kevloral
[14:50] <toad_> http://amphibian.dyndns.org/simulations3-400x25-avg25.png
[14:50] <toad_> promising imho
[14:51] <toad_> hmmm
[14:51] <toad_> terrence is giving me problems uploading the new jar
[14:52] <toad_> cat2.ath.cx...
[14:53] <toad_> mikeeusa here?
[14:55] <hackeron> what browser would you recommend for freenet?
[14:55] <toad_> firefox
[14:55] <toad_> probably
[14:55] <hackeron> toad_: with that network configuration extension?
[14:55] <toad_> NOT Internet Explorer (it has some specific problems with freenet)
[14:55] <toad_> hackeron: it's not an extension
[14:56] <toad_> just run about:config
[14:56] <toad_> and change the relevant settings
[14:56] <toad_> Opera.. well if you know what you're doing and turn off the mime type autodetection
[14:56] <hackeron> toad_: PS, I dont use windows.
[14:56] <toad_> konqueror.. i don't use it much for freenet... is it secure? it always seemed quite objecty which means it might not be safe..
[14:57] <toad_> objecty/plugin-happy
[14:57] <hackeron> toad_: not anymore than firefox
[14:57] <hackeron> toad_: err, there's quite a lot of things I can change in firefox, can you be more specific?
[14:57] <toad_> cool
[14:57] <toad_> hackeron: look for network.http
[14:57] <toad_> there are various connection settings
[14:58] <toad_> oooh http://amphibian.dyndns.org/simulations3-400x25-avg25.png
[14:59] <TheSeeker> I was able to change the connections limit in IE, though I understand that there's other potential security risks, I don't think I've viewed anything on Freenet that would get me in more trouble than what I use the regular net for on a regular basis.
[14:59] <toad_> seems quite jumpy.. odd..
[14:59] <toad_> TheSeeker: well that's your risk
[14:59] <hackeron> toad_: what are those graphs btw?
[14:59] <toad_> hackeron: simulation results
[15:00] <hackeron> toad_: of?
[15:00] <toad_> hackeron: we're trying to put freenet on a slightly more rigorous basis
[15:00] <toad_> one mode of attack in this endeavour is to simulate freenet networks
[15:00] <toad_> and try to show that freenet can a) work and b) scale, at least in idealized theoretical conditions
[15:00] <toad_> this has not been done yet for NG routing
[15:00] <toad_> but i am doing it
[15:01] <toad_> also it lets you simulate all the little routing changes we implement
[15:01] <toad_> and find which ones are likely to be useful
[15:02] <toad_> X axis is number of (started) requests. Y axis is probability of success of a given request
[15:02] <toad_> green is the simulation without connection churn i.e. 25 random connections set on startup, which never change
[15:02] <toad_> red is with LRU and the freenet connection/path-folding algorithm (connect to the datasource, if you can, if you have enough data to drop a node, etc)
[15:06] <hackeron> toad_: hmm, so it scales up quite nicely then?
[15:07] <toad_> hackeron: that's what we are trying to find out
[15:07] * sdogi (~java@84-50-18-247-dsl.kjj.estpak.ee) has joined #freenet
[15:09] <hackeron> toad_: well, looking good so far :)
[15:09] <toad_> updated... blue is with connection churn but with 13 HTL
[15:09] <toad_> this is in case red falls down
[15:09] <toad_> which it may well
[15:10] <toad_> hackeron: green isn't
[15:10] <toad_> red may be
[15:10] <toad_> i'm only really comfortable if it keeps on going, reaches 99% and stays there :)
[15:10] <hackeron> anything I can do to make freenet eat less ram? :( -- its eating 31% atm.
[15:10] <toad_> like the 100x25's did
[15:10] <toad_> hackeron: hmmm
[15:10] <toad_> not sure
[15:11] <toad_> what does top say? are you confusing RSS with the other one?
[15:11] <hackeron> 14036 freenet 20 0 531m 151m 39m S 0.0 31.0 0:29.51 java
[15:12] <toad_> i suggest the 151m is the real figure
[15:12] <toad_> what are the field names? 531m is VIRT, right?
[15:12] <toad_> 151 is RSS?
[15:12] <toad_> 39 is shared?
[15:12] <toad_> virtual isn't very meaningful iirc
[15:12] <hackeron> 151 is RES
[15:12] <hackeron> 151 out of 512
[15:12] <toad_> right
[15:12] <hackeron> thats FAR too much
[15:12] <toad_> hmmm
[15:12] <toad_> well it's quite heavy yeah
[15:13] <hackeron> should use at most 30mb
[15:13] <toad_> you can try reducing the -Xmx, or the thread limit, or both
[15:13] <toad_> there's a few things you can do
[15:13] <toad_> but it's not a priority at the moment, because ram is cheap, working code is expensive, working decentralized anonymous p2p is hard, and freedom is usually bought with human blood
[15:14] <toad_> we'll optimize when it works :)
[15:14] <toad_> bbiab, going to get food
[15:14] <toad_> the graphs are auto-updated for anyone interested
[15:14] <toad_> at least this one is
[15:14] <toad_> for now :)
[15:15] * sdogi_ (~java@84-50-17-178-dsl.kjj.estpak.ee) Quit (Connection timed out)
[15:15] <hackeron> it doesnt work? :p
[15:15] <toad_> it doesn't work anything like as well as we hope it will one day
[15:15] <toad_> :)
[15:16] <hackeron> hmm, I see :)
[15:16] <toad_> of course it's useful to many people, but it's still very much alpha, maybe even pre alpha
[15:16] <toad_> bbiab
[15:16] <hackeron> I thought it was a stable version already, hmmm
[15:16] <hackeron> how many years is it in development?
[15:26] * Hory (~Miranda@82.76.81.56) Quit (Connection timed out)
[15:29] <toad_> lots :)
[15:30] <hackeron> and still in alpha, eh? :p -- think its close to v1.0, lol?
[15:32] <TheSeeker> 0.5 is close to 1.0?
[15:32] <greycat> I wouldn't say so.
[15:32] <TheSeeker> 0.6 may be fairly nearby, but I don't think anythign is even really planned past that point...
[15:33] <hackeron> TheSeeker: fairly near by? -- next month?
[15:33] <greycat> ask the amphibious one
[15:34] * i2p_iip (~changate@host.teitel.net) Quit (Remote closed the connection)
[15:34] <TheSeeker> "when it's ready" ?
[15:35] <TheSeeker> I think routing needs to be proven to work first, then bugs squashed, then a network reset and testing on 'stable' ... Then work on 0.6 features can start? :P
[15:37] <hackeron> so between never and soon, its closer to never, eh? :p
[15:40] <TheSeeker> nah, sims are likely to help identify problems, and what's capable of working and scaling well...
[15:54] <hackeron> now freenet is using 204mb ram :(
[15:55] <hackeron> or 42%
[15:56] <hackeron> and growing :(
[16:00] * Hory (~Miranda@82.76.81.56) has joined #FreeNet
[16:06] * pittaman (~mistery_b@d51A410CA.kabel.telenet.be) Quit (Client Quit)
[16:14] * Hory (~Miranda@82.76.81.56) Quit ("bbabl")
[16:20] * sanity (~ian@81-178-106-165.dsl.pipex.com) Quit ()
[16:27] <TheSeeker> freenet will use as much RAM as you let it.
[16:28] <hackeron> TheSeeker: how do I limit it?
[16:30] * Elly (~Elly@ool-182c3b26.dyn.optonline.net) has joined #freenet
[16:35] <sdogi> TheSeeker: and if you don't let it use much it will perform badly
[16:36] * orange_ (~orange@pm6-30.bahnhof.se) has joined #freenet
[16:39] <TheSeeker> there's an option in the configurato, or you can set it in flaunch.ini in windows ... under linux I think you have to add an -Xmx###M to your java line...
[16:40] <sdogi> i wish entropy were better
[16:40] <sdogi> it at least runs on my 486
[16:43] * greycat (~wooledg@192.35.79.70) Quit ("This time the bullet cold rocked ya / A yellow ribbon instead of a swastika")
[16:44] * Hory (~Miranda@82.76.81.56) has joined #FreeNet
[16:45] <hackeron> sdogi: entropy is really very clean. But the mainlist lists say the author abandoned the project and no one seems to care enough to fork it :(
[16:46] <toad_> yes, but if you set Xmx too low, it'll crash...
[16:46] <toad_> other option is to reduce # threads, for example
[16:47] <sdogi> hackeron: damn
[16:48] <hackeron> python has some really fantastic network programming libraries. Wonder if something like this can be written in python and be efficient enough -- afterall, bittorrent is python :)
[16:53] <sdogi> don't know about that
[17:32] * sanity (~ian@81-178-106-165.dsl.pipex.com) has joined #freenet
[17:38] * sanity (~ian@81-178-106-165.dsl.pipex.com) has left #freenet
[17:40] * jay (~jcl@198.38.10.200) Quit ()
[17:55] * Hadaka (naked@naked.iki.fi) Quit (Remote closed the connection)
[17:55] * Naked (naked@naked.iki.fi) has joined #freenet
[17:56] * Naked is now known as Hadaka
[18:03] <toad_> http://amphibian.dyndns.org/simulations3-400x25-avg25.png
[18:03] <toad_> yay!
[18:04] <toad_> hackeron: maybe, but lets get it working first
[18:06] <toad_> well into the 90s with 13 HTL, and 10 isn't far off
[18:06] <toad_> may well make it with more time
[18:06] <toad_> i don't know if it'll stay there though
[18:07] <toad_> need to run it out to 6-10M requests
[18:25] <hackeron> toad_: no, what I mean is you can prototype it in python to save conciderable time to get it working, then rewrite the parts that need speed in java :) -- have a look at jython.
[18:28] * Elly is now known as Elly|afk
[18:43] * Ash-Fox (HAL-9000@acw109.neoplus.adsl.tpnet.pl) Quit (Nick collision from services.)
[18:43] * Ash-Fox (HAL-9000@adf142.neoplus.adsl.tpnet.pl) has joined #FreeNET
[18:49] * Zer0Her0 (~Z@17.255.240.70) Quit ()
[18:52] <toad_> hackeron: heh
[18:52] <toad_> hackeron: that's what i have to say :)
[18:52] <hackeron> toad_: sorry, not a fan of java :p
[18:52] <toad_> you can't prototype it in python after you've already got 130kloc of "working" java code
[18:53] <toad_> you can go back to day one, of course
[18:53] * TrigZ (~Opposite@203-173-20-43.dyn.iinet.net.au) has joined #freenet
[18:54] <hackeron> toad_: well, it was pretty much designing the system. Its not bigger than openoffice and openoffice is getting rewritten as do many other projects :)
[18:54] <toad_> "java is bloated" "no its not" "yes it is, see benchmarks" "no it's not, see other benchmarks"... well of course it's bloated, although speed wise it's reasonable (except on amd64 for some strange reason), and it's controlled by sun (but not once we liberate it with classpath/kaffe/gcj)
[18:54] <toad_> hackeron: in a completely different language? that'll take a while...
[18:54] * TrigZ (~Opposite@203-173-20-43.dyn.iinet.net.au) Quit (Client Quit)
[18:55] <hackeron> toad_: just think of jython as java for dummies :p -- check out the twisted web development framework for it.
[18:55] <toad_> hackeron: anyway, there is no grounds at present for rewriting from scratch; it might improve some aspects of local node performance, but the key stuff is all network level
[18:55] <toad_> and we can't afford to sacrifice a year plus's work rewriting it in another language
[18:56] * TrigZ (~Opposite@203-173-20-43.dyn.iinet.net.au) has joined #freenet
[18:56] <toad_> http://amphibian.dyndns.org/simulations3-400x25-avg25.png
[18:56] <hackeron> toad_: you think it will really take longer than a month (tops)?
[18:56] <toad_> blue turning downwards?
[18:56] <toad_> hackeron: i don't even know python. many of the devs don't. and yes, 130kloc cannot be rewritten and debugged in ANY language in a month.
[18:57] <toad_> anyway we'd have to agree on a language
[18:57] <toad_> and that'll never happen
[18:57] <toad_> :)
[18:57] <hackeron> toad_: yeah, silly idea, but python still rocks :)
[18:58] <toad_> perhaps it does
[18:59] <hackeron> PS, who chose java for this?
[18:59] <toad_> ian
[18:59] <toad_> i won't state an opinion on java because that would start a flamewar and i'm going to bed...
[19:00] <hackeron> toad_: heh, I love the good old java bashing :)
[19:00] <hackeron> toad_: just a taster, its a bad opinion, right?
[19:03] <toad_> okay, here's my position: we have to use java right now, as i have explained. java is bad because it is non-free (this isn't just a political problem); this is being rectified, and we should spend some effort to help out. i don't actually know many languages - java, c, c++, a little perl, 8088 assembly, shell script, gw-basic/qbasic/wordbasic. but i see no particular reason to hate java. interfaces are a nice simplification that avoid
[19:03] <toad_> however i'm not a conosseur of these things
[19:03] <toad_> as i said, i don't know python
[19:03] <toad_> i don't know lisp
[19:04] <toad_> i know a little forth, as well as the above :)
[19:04] <toad_> i accept that C++ is a bit of a monster, but it has its own particular appeal even so
[19:05] <toad_> someday i'll learn python and probably like it a lot
[19:05] <hackeron> toad_: yeah, a network designed for free speech thats programmed on a closed language, interesting :)
[19:05] <toad_> java is essentially c++ with the nasty bits taken out
[19:05] <toad_> with a reasonable library, and some cross platform support
[19:05] <toad_> hackeron: yeah, it's a problem for packaging as well as politics
[19:06] <toad_> hackeron: it has been worked on; it will be worked on further
[19:06] <hackeron> toad_: its developer friendly, but all java apps I tried run like crap
[19:06] <toad_> hackeron: too much memory, or what?
[19:06] <toad_> have you tried the flight simulator?
[19:06] <hackeron> toad_: they just dont feel like native apps
[19:06] * toad_ never got that to work but it sounded really impressive
[19:07] <toad_> hackeron: whereas python is as responsive as C?
[19:07] <hackeron> toad_: for example eclipse or whatever it was called, or sancho for mldonkey, the java parts of openoffice, you name it -- it all just doesnt feel right, runs slow, eats memory, etc.
[19:07] <sdogi> not really
[19:07] <hackeron> toad_: python is crawling compared to java :)
[19:07] * sdogi hates memory eaters too
[19:07] <toad_> the precompiled (GCJ) eclipse loads a LOT faster
[19:07] <toad_> and I've explained my views on memory usage
[19:07] <hackeron> toad_: but it reduces the code amount for a factor of 10
[19:07] <toad_> if you require me to repeat them i'll scan for the quote
[19:08] <hackeron> toad_: "ram is cheap" remark?
[19:08] <toad_> yeah
[19:08] <hackeron> toad_: hard drive is cheaper :p
[19:08] <toad_> so?
[19:09] <sdogi> ram is cheap is like an excuse
[19:09] <toad_> hackeron: well i'll believe that when i see it
[19:09] <toad_> sdogi: no, it's a matter of priorities
[19:09] <sdogi> when i would use all my java programs once i would need at least 2 gigs of ram
[19:09] <sdogi> also it would make everything slow
[19:09] <toad_> either we work on making a single freenet node fit into 640kB and run on an IBM XT, or we work on making the network as a whole work
[19:10] <sdogi> because loading from ram is pretty slow, specially when there is gigabyte of data to load
[19:10] <toad_> I would argue that beyond a certain point, the latter is far more important than the former
[19:10] <sdogi> although i can totally understand your problems
[19:10] <toad_> obviously there is a point at which you have to address memory usage problems
[19:10] <toad_> e.g. if they exclude large numbers of potential users
[19:10] <toad_> but i don't think it does
[19:10] <hackeron> toad_: well, looking at entropy -- it just feels native, runs fast and takes a total of 30Mb ram for 100 threads.
[19:11] <toad_> entropy is not comparable to freenet because it doesn't tackle most of the hard problems
[19:11] <toad_> it just has a fixed, 4-bit (!!!) specialization for each node
[19:11] <sdogi> yeah, but entropy seems like non working product sadly :(
[19:11] <toad_> it is probably far more attackable than freenet
[19:11] <toad_> and it certainly doesn't scale as well as freenet does
[19:12] <hackeron> toad_: it doesnt? -- lets see some numbers :p
[19:12] <toad_> hackeron: how far can it possibly scale with only 4 bits of specialization?
[19:12] <toad_> 2^4 = 16
[19:12] <hackeron> toad_: why wouldnt it scale up if its not limited by hardware resources? -- freenet stops working on 1ghz, 512Mb ram if I have more than 15 tabs open because of lack of CPU power.
[19:12] <toad_> it won't scale because the network architecture doesn't scale
[19:13] <toad_> you can implement something that works in virtually no RAM on one PC, on a network of 10 nodes
[19:13] <hackeron> toad_: what do you mean by 16?
[19:13] <toad_> and it be completely useless on a network of 10,000 nodes
[19:13] <toad_> freenet's speed problems are not fundamentally down to java, nor are they down to general slowness of the code, or even memory usage
[19:13] <toad_> they are down to the fact that we haven't got the routing algorithm pinned down yet
[19:15] <toad_> hackeron: afaics, entropy nodes go into 16 "buckets", and each node wants to be connected to at least one of each bucket; routing decisions are based only on which bucket the key is in, and perhaps some key-independant performance data
[19:15] <toad_> the result of this is that it won't scale much beyond 16*<acceptable broadcast htl> nodes
[19:15] <toad_> acceptable broadcast htl is probably something like 40 at most
[19:15] <toad_> s/afaics/iirc
[19:15] <hackeron> toad_: hmmm
[19:16] <hackeron> toad_: that sounds really bad. No wonder he dropped out of the project.
[19:16] <toad_> but i don't really know how it works
[19:16] <toad_> i may be completely wrong
[19:16] <toad_> and even the stupidest projects have their uses sometimes e.g. the p2p system that anonymizes by ip spoofing
[19:17] <toad_> i don't remember what it was called
[19:17] <toad_> anyway, the hard problems are on the network architecture level
[19:17] <toad_> that's the point i'm trying to make
[19:18] <toad_> and now, i go to get supper
[19:27] <hackeron> toad_: why do I just get a blank page from time to time?
[19:28] <hackeron> toad_: and my freenet is taking 100% cpu at almost all times
[19:30] <hackeron> toad_: even in middle of download things refresh to a blank page when under load.
[19:32] <hackeron> bah: 14036 freenet 20 0 550m 261m 39m S 0.0 53.7 0:29.51 java
[19:37] <hackeron> toad_: PS, this is a recent description of what you are trying to do? http://freenetproject.org/index.php?page=ngrouting
[19:41] * kevloral is now known as kevlAway
[19:48] <toad_> hackeron: yes
[19:48] <toad_> hackeron: possibly it is still opening connections?
[19:48] <toad_> what's your cpu, and your node uptime?
[19:51] <hackeron> toad_: 1ghz, still 100% since my last message - uptime 5 hours 26 minutes
[19:54] * Hory (~Miranda@82.76.81.56) Quit ("CyberLore.net - Recommendations on the best games, freeware and websites.")
[19:54] <hackeron> toad_: its absolutely brutal on resources. Opening local pages like downloads and open connections takes up to and over a minute
[19:55] <hackeron> toad_: no wonder its slow in routing
[19:55] <hackeron> 361 MiB transferred total though, quite a bit
[19:56] <hackeron> 11 active downloads -- all keep failing or refreshing to blank page.
[19:57] <hackeron> so I think having 600 active downloads like mldonkey is out of the question :p
[20:00] <hackeron> works painfully slow, but works for 2 file downloads -- but doesnt scale up at all for single user :(
[20:08] * Elly|afk is now known as Elly
[20:14] * jay (jcl@ool-18bf6dac.dyn.optonline.net) has joined #freenet
[20:25] <toad_> hackeron: most of the time it uses a few percent cpu usage according to what i have heard
[20:25] <toad_> obviously it uses more for me because i run with excessive logging
[20:26] <toad_> if you are running it on a 486 that would explain it
[20:26] <toad_> it might be a bootstrapping problem alternatively
[20:26] <toad_> as regards mldonkey... you mean 600 QUEUED downloads, right? :)
[20:26] <toad_> how is that different to freenet DNF'ing?
[20:27] <toad_> also, a key determinator of performance is the number of connections you have; how many connections do you have? (open connections)
[20:28] * Sandman (~s@CPE-24-164-232-98.new.rr.com) has joined #freenet
[20:28] <toad_> I have excessive logging and it works a lot better than you describe despite that
[20:29] <toad_> so it's probably a problem on your end
[20:29] <toad_> anyway now i'm going to bed
[20:38] * Hadaka (naked@naked.iki.fi) Quit (Read error: 110 (Connection timed out))
[20:38] * jay (jcl@ool-18bf6dac.dyn.optonline.net) Quit ("Client exiting")
[20:42] * Naked (naked@naked.iki.fi) has joined #freenet
[20:42] * Naked is now known as Hadaka
[20:52] <hackeron> toad_: 600 queued, 300 active
[20:52] <hackeron> toad_: and its a Via C3-2 1ghz
[20:53] <hackeron> the nehemiah one.. not the crappy c3
[21:03] <leexx> n:)
[21:06] <leexx> lol 600 qued
[21:06] <leexx> emule realy needs to change the way the ratio system works
[21:07] <leexx> Added source to bad source list (Global) - file ??? : 213.65.64.169 'joshuasaxxx' (MLdonkey v0.42.4,Error/None)
[21:07] <leexx> whats mldonkeys web site
[21:12] * leexx is now known as lexx-zzz
[21:13] * orange_ (~orange@pm6-30.bahnhof.se) Quit ()
[21:15] * Zer0HiT (~Z@alb-24-92-54-145.nycap.rr.com) has joined #freenet
[21:16] * ike (wangcity@221.232.191.209) has left #freenet
[21:21] * i2p_iip (~changate@host.teitel.net) has joined #freenet
[21:33] * mazzanet (~mazzanet@mazzanet.user) Quit (Read error: 54 (Connection reset by peer))
[21:35] * mazzanet (~mazzanet@203-217-32-152.dyn.iinet.net.au) has joined #freenet
[21:35] <mikeeusa> hi
[21:36] * Bonkers (bonker@dyn-wien-209-200.dyn.columbia.edu) Quit ("Client exiting")
[21:44] * guido^pe (~unknown@dsl-213-023-238-133.arcor-ip.net) Quit (Remote closed the connection)
[21:59] * jay (jay@ool-18bf6dac.dyn.optonline.net) has joined #freenet
[22:06] * Elly is now known as Elly|afk
[22:30] <TheSeeker> http://www.google.com/search?q=mldonkey ?
[22:48] <mikeeusa> yo
[22:48] <mikeeusa> Torrent sites got arrowed today
[22:48] <mikeeusa> finnish police* raided them
[22:49] <mikeeusa> *police are people who breakup families, put men in jail at their wive's will, and are generally a bad thing
[22:49] <mikeeusa> (guns can replace police in any society)
[22:51] <mikeeusa> so perhapse freenet will get more popular?
[22:56] * linagee (~linagee@netblock-66-245-227-51.dslextreme.com) Quit (Client Quit)
[22:56] * linagee (~linagee@netblock-66-245-227-51.dslextreme.com) has joined #freenet
[23:09] * Elly|afk (~Elly@ool-182c3b26.dyn.optonline.net) Quit (Read error: 110 (Connection timed out))
[23:21] * linagee (~linagee@netblock-66-245-227-51.dslextreme.com) Quit (Client Quit)
[23:23] * matahar1 (~matahari@queeg.cs.mu.OZ.AU) has joined #freenet
[23:41] * jay (jay@ool-18bf6dac.dyn.optonline.net) Quit ("Leaving")
These logs were automatically created by Jay Oliveri with his gimp hapi on irc.freenode.net.
