i2p.www/www.i2p2/pages/meeting190.html

{% extends "_layout.html" %}
{% block title %}I2P Development Meeting 190{% endblock %}
{% block content %}<h3>I2P dev meeting, November 21, 2006</h3>
<div class="irclog">
<p>15:02 &lt; jrandom&gt; 0) hi</p>
<p>15:02 &lt; jrandom&gt; 1) Net status</p>
<p>15:02 &lt; jrandom&gt; 2) Syndie dev status</p>
<p>15:02 &lt; jrandom&gt; 3) ???</p>
<p>15:02 &lt; jrandom&gt; 0) hi</p>
<p>15:02  * jrandom waves</p>
<p>15:02 &lt; jrandom&gt; weekly status notes posted up at http://dev.i2p.net/pipermail/i2p/2006-November/001319.html</p>
<p>15:03 &lt; jrandom&gt; since that one is pretty short, lets jump on in to 1) net status</p>
<p>15:04 &lt; jrandom&gt; things are looking pretty good atm, network seems pretty steady</p>
<p>15:04 &lt;+zzz&gt; I invented a "peer capacity index"</p>
<p>15:04 &lt;+zzz&gt; on the dashboard...</p>
<p>15:04 &lt;+zzz&gt; so far not sure it is helpful though</p>
<p>15:04 &lt; jrandom&gt; ah yeah, sorry, metioned that one last week - looks quite useful, thanks!</p>
<p>15:05 &lt; jrandom&gt; interesting to see the disparity out there so clarly</p>
<p>15:05 &lt;+zzz&gt; the idea is the ratio of high-cap routers to low-cap routers, which is obviously important to tunnel build %</p>
<p>15:06 &lt;+zzz&gt; I'm removing routers from stats that I don't get a netdb update for in 1.5 hours but that seems too quick, I think it is skewing the stats</p>
<p>15:07 &lt; jrandom&gt; ah, ok, that would explain it.  are you still harvesting?</p>
<p>15:07 &lt; jrandom&gt; (or wget'ing from dev.i2p.net?)</p>
<p>15:08 &lt;+zzz&gt; yes</p>
<p>15:08 &lt; jrandom&gt; cool</p>
<p>15:08 &lt;+zzz&gt; netDb.harvestDirectly=false</p>
<p>15:08 &lt;+zzz&gt;   netDb.shouldHarvest=true, right?</p>
<p>15:09 &lt; jrandom&gt; so the stats we've had before were largely based on routers that were so bad the user shut them down & disapeared then?</p>
<p>15:09 &lt; jrandom&gt; right</p>
<p>15:10 &lt;+zzz&gt; it's always been 1.5 hours, but plotting the M/N/O routers, they seem to come and go when intuitively they should stay pretty constant</p>
<p>15:10 &lt; jrandom&gt; ah ok</p>
<p>15:10 &lt;+zzz&gt; you can see spikes/dips in all the data that last 1.5 hours :)</p>
<p>15:11 &lt; spaetz&gt; net seems pretty stable. Yep</p>
<p>15:12 &lt;+zzz&gt; thats all I have for that topic</p>
<p>15:12 &lt; spaetz&gt; I'd like to know if jrandom completely focuses on syndie nowadays or if he still looks at i2p dev.</p>
<p>15:12 &lt; spaetz&gt; or if this is just a bit on hte backburner temporarily</p>
<p>15:13  * jrandom completely focuses on syndie nowadays, but will work on i2p both when there are problems and once syndie is established</p>
<p>15:13  * spaetz thanks for the information</p>
<p>15:14  * spaetz is fine with this</p>
<p>15:15 &lt; jrandom&gt; w3wt.  yeah, steadystate means syndie dev can continue, but if there are problems, of course i reprioritize</p>
<p>15:15 &lt; jrandom&gt; ok, anyone have anything else on 1) net status?</p>
<p>15:15 &lt; Walter&gt; I have a random question.</p>
<p>15:15 &lt; jrandom&gt; hit me Walter </p>
<p>15:17 &lt; Walter&gt; Assume you have 100Mb/s BW, what kind of server would you need to saturate it as an I2P node?</p>
<p>15:17 &lt; jrandom&gt; doesnt matter</p>
<p>15:17 &lt; jrandom&gt; i2p does not and will not saturate 100Mbps</p>
<p>15:18 &lt; Walter&gt; Assume one wanted to make use of available BW.</p>
<p>15:18 &lt; jrandom&gt; you would not.</p>
<p>15:19 &lt; spaetz&gt; I've got 150kbs up and down and it uses like 25% of a vserver (Dell shared with a dozen others)</p>
<p>15:19 &lt; jrandom&gt; that exceeds the capacity of the entire network</p>
<p>15:19 &lt; spaetz&gt; 25%CPU that is</p>
<p>15:19  * spaetz admits that's not really a precise answer and shuts up</p>
<p>15:20 &lt; jrandom&gt; the routers themselves have a mem v. throughput tradeoff, making it less likely that a router can even push &gt; 3-350KBps</p>
<p>15:20 &lt; jrandom&gt; (of course, that tradeoff can be tweake to allow higher rates, but thats not an issue)</p>
<p>15:21 &lt; jrandom&gt; using bandwidth is *BAD* unless that bandwidth is being used only when necessary</p>
<p>15:22 &lt;+zzz&gt; the network is averaging about 1.5 MBps (=12 Mbps) total traffic over the last 3 months</p>
<p>15:23 &lt; Walter&gt; I see.</p>
<p>15:24 &lt;+fox&gt; &lt;LeerokKitchen&gt; Field trip!</p>
<p>15:26 &lt; jrandom&gt; ok, if there's nothing else for 1) net status, lets jump on over to 2) syndie dev status</p>
<p>15:26 &lt; jrandom&gt; progress here continues, and i've been doing testing both on windows and linux</p>
<p>15:28 &lt; jrandom&gt; current battle is on the forum management interface, though since the text interface is already embedded, all functionality is already in place</p>
<p>15:29 &lt; jrandom&gt; not much more news to discuss on that front though</p>
<p>15:30 &lt; jrandom&gt; anyone have any questions/comments/concerns on 2) syndie dev status?</p>
<p>15:33 &lt; jrandom&gt; ok, lets jump on to 3) ???</p>
<p>15:33 &lt; jrandom&gt; y'all have anything else for the meeting?</p>
<p>15:34 &lt;+fox&gt; &lt;blx&gt; when will gpl java be usable with i2p=</p>
<p>15:34 &lt;+fox&gt; &lt;blx&gt; ?</p>
<p>15:35 &lt; Complication3&gt; I guess it depends on when gpl java will be usable on various distros</p>
<p>15:35 &lt; Complication3&gt; Or available for download from Sun</p>
<p>15:36 &lt; Complication3&gt; But it feels like a moot point, since it's the same Java which is usable already now</p>
<p>15:36 &lt; Complication3&gt; GPL would only let it be packaged more conveniently, and improved upon</p>
<p>15:37 &lt; jrandom&gt; (and i2p already works with gcj/kaffe, though not all of the client apps)</p>
<p>15:37  * Complication3 quickly reads backlog</p>
<p>15:37 &lt; jrandom&gt; ((and syndie works with gcj/kaffe completely))</p>
<p>15:38 &lt;+fox&gt; &lt;blx&gt; Compilation, thats what they want you to think ;)</p>
<p>15:38 &lt;+fox&gt; &lt;blx&gt; but ok, i got my question answered.</p>
<p>15:38 &lt;+fox&gt; &lt;blx&gt; Complication even. misread</p>
<p>15:39 &lt; Complication3&gt; blx: well, the sources are available already now, it's just that few read and compile them</p>
<p>15:39 &lt; jrandom&gt; (and you can even modify and use those modifications, you just can't distribute your mods)</p>
<p>15:40 &lt; koff&gt; when will i2p have the logging functionality suggested by the proposed laws i heard about?</p>
<p>15:41 &lt; jrandom&gt; never</p>
<p>15:41 &lt;+zzz&gt; hahahaha</p>
<p>15:41  * Complication3 suspects never :)</p>
<p>15:41 &lt;+fox&gt; &lt;blx&gt; what laws?</p>
<p>15:41  * jrandom assumes you refer to .de/.eu data retention issues</p>
<p>15:41 &lt; Complication3&gt; Someone in the forum talked of a (proposed) law in Germany</p>
<p>15:42 &lt; jrandom&gt; (and then the .us ones in a few years)</p>
<p>15:42 &lt; Complication3&gt; They could have spelled it out better though</p>
<p>15:42 &lt; jrandom&gt; aye, 'tis just proposed, but not a big suprise</p>
<p>15:43 &lt; Complication3&gt; I personally think: it's not like data retention laws aren't being broken left and right already</p>
<p>15:43 &lt; Complication3&gt; Breaking a dozen more of them? I personally wouldn't care much...</p>
<p>15:44 &lt; Complication3&gt; In short, I want to see how they're going to enforce it</p>
<p>15:44 &lt; tea&gt; like they did with napster : arrest everyone</p>
<p>15:45 &lt; Complication3&gt; If they manage to make a good try, something will need to be found to thwart that ("not in my country" peering principle for countries where insanity prevails)</p>
<p>15:45 &lt;+fox&gt; &lt;LeerokLacerta&gt; That reminds me of a song.</p>
<p>15:45 &lt;+fox&gt; &lt;LeerokLacerta&gt; http://2ch.ru/mu/src/1163070550597.mp3</p>
<p>15:46 &lt; tea&gt; turning all data traffic over to anonymous networks might help ...</p>
<p>15:47 &lt; Complication3&gt; Just ignoring them en masse has worked for plain ordinary pirates...</p>
<p>15:47 &lt; Complication3&gt; You can arrest one person ignoring you. Can't do that with several hundred thousand.</p>
<p>15:47 &lt; tea&gt; that's no argument for a german :)</p>
<p>15:47 &lt;+fox&gt; &lt;modulus&gt; you can</p>
<p>15:47 &lt;+fox&gt; &lt;modulus&gt; hitler did</p>
<p>15:48 &lt; Complication3&gt; That's only because nobody bothered removing him</p>
<p>15:48 &lt; jrandom&gt; *cough*</p>
<p>15:48 &lt; Complication3&gt; Had they taken up arms, it wouldn't have worked</p>
<p>15:48 &lt; Complication3&gt; (sorry, far off topic, yes)</p>
<p>15:48 &lt; tea&gt; still, one does feel important in being paranoid</p>
<p>15:48 &lt;+fox&gt; &lt;modulus&gt; that said i think i2p could comply with data retention laws without damaging anonimity, but there's no reason to do that.</p>
<p>15:48 &lt; jrandom&gt; ok, well, i think we've addressed the i2p-related issue there ;)</p>
<p>15:48 &lt; tea&gt; sry</p>
<p>15:49 &lt; jrandom&gt; aye modulus</p>
<p>15:49 &lt; jrandom&gt; (we already assume individual users are logging everything anyway, as are the isps)</p>
<p>15:49 &lt;+fox&gt; &lt;modulus&gt; right, so a DR-enabled i2p wouldn't be the end of the world</p>
<p>15:51 &lt; Complication3&gt; Someone would have to bother forking that, though... :P</p>
<p>15:52  * jrandom keeps my mouth shut ;)</p>
<p>15:52 &lt; jrandom&gt; ok, anyone have anything else for the meeting?</p>
<p>15:53 &lt; jrandom&gt; if not</p>
<p>15:53  * jrandom winds up</p>
<p>15:53  * jrandom *baf*s the meeting closed</p>
</div>
{% endblock %}