79 lines
3.0 KiB
HTML
79 lines
3.0 KiB
HTML
|
{% extends "_layout.html" %}
|
||
|
|
{% block title %}I2P Status Notes for 2005-12-13{% endblock %}
|
||
|
|
{% block content %}<h3>I2P Status Notes for 2005-12-13</h3>
|
||
|
|
<pre>-----BEGIN PGP SIGNED MESSAGE-----
|
||
|
|
Hash: SHA1
|
||
|
|
|
||
|
|
Hi y'all, 'tis that time again
|
||
|
|
|
||
|
|
* Index
|
||
|
|
1) Net status and load testing
|
||
|
|
2) I2PSnark
|
||
|
|
3) Syndie
|
||
|
|
4) ???
|
||
|
|
|
||
|
|
* 1) Net status and load testing
|
||
|
|
|
||
|
|
Not much has changed in the last week, though we have had a little
|
||
|
|
trouble on the irc servers. Most of my time has been spent on
|
||
|
|
various network load tests, trying to track down the source of our
|
||
|
|
current throughput bottleneck. Tunnels themselves can transfer at
|
||
|
|
a high rate. SSU can transfer at an even higher rate. And the
|
||
|
|
streaming lib can push even more data than that. There are two
|
||
|
|
general possibilities for the bottleneck - some software bug, or
|
||
|
|
suboptimal (aka bad) peer selection. What I'm doing now is trying
|
||
|
|
to determine which it is. Preliminary indications are that some
|
||
|
|
peers (perhaps 25%) have no trouble pushing 20-100KBps through a
|
||
|
|
tunnel, while the rest stay in the 1-4KBps range under load.
|
||
|
|
However, thats only very preliminary results, and more realistic
|
||
|
|
load tests are ongoing.
|
||
|
|
|
||
|
|
We've also got zzz's latest changes improving the tunnel creation in
|
||
|
|
CVS, and reports from the field have been good - nice work zzz! I'm
|
||
|
|
not sure when 0.6.1.8 will come out, but if not by the weekend, then
|
||
|
|
early next week. We'll see.
|
||
|
|
|
||
|
|
* 2) I2PSnark
|
||
|
|
|
||
|
|
The other day while gathering data from some load tests I patched up
|
||
|
|
I2PSnark so that it can deal with torrents created by Azureus and
|
||
|
|
I2PRufus again. The problem was that it didn't know how to deal
|
||
|
|
with some of the newer torrent attributes ("name.utf-8" and
|
||
|
|
"path.utf-8"), which was causing it to mess up the infohash (and
|
||
|
|
hence be unable to talk to the tracker or any peers). Thats fixed
|
||
|
|
in CVS, so it should be a workable anonymous bt client again - if
|
||
|
|
anyone has problems getting it to deal with a torrent that I2P-BT
|
||
|
|
can deal with, please let me know.
|
||
|
|
|
||
|
|
* 3) Syndie
|
||
|
|
|
||
|
|
Polecat has been hacking away, with the latest addition offering
|
||
|
|
thumbnail support for attachments in SML (see the smlref.jsp on the
|
||
|
|
current CVS build for more details). I also put together a first
|
||
|
|
pass at the blog interface, but its not really that great, so
|
||
|
|
another pass is definitely in the offing. Keep an eye on your
|
||
|
|
Syndie instance for the latest Syndie news, of course.
|
||
|
|
|
||
|
|
* 4) ???
|
||
|
|
|
||
|
|
There's lots going on, as always, though my focus still continues to
|
||
|
|
be on improving performance to the point where its adequate for
|
||
|
|
reasonable uses. As mentioned on the forum, that I2Phex sourceforge
|
||
|
|
fiasco is finally taken care of, and discussion continues regarding
|
||
|
|
jbigi and jcpuid builds and dependencies (more news on that front
|
||
|
|
when there's more to say). Anyway, as always swing on by the
|
||
|
|
meeting tonight at 20:00 UTC and say hey!
|
||
|
|
|
||
|
|
=jr
|
||
|
|
-----BEGIN PGP SIGNATURE-----
|
||
|
|
Version: GnuPG v1.4.1 (GNU/Linux)
|
||
|
|
|
||
|
|
iD8DBQFDnyJJWYfZ3rPnHH0RAn2eAJ0YmiNTCGm46zkyH44YB/gJ1tysOgCfX5UU
|
||
|
|
RgvenKVDEEUbUMi8ePPOkuk=
|
||
|
|
=GZ6w
|
||
|
|
-----END PGP SIGNATURE-----
|
||
|
|
|
||
|
|
|
||
|
|
</pre>
|
||
|
|
{% endblock %}
|