[i2p] weekly status notes [nov 16]

[jrandom]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi y'all, its tuesday again

* Index
1) Congestion
2) Streaming
3) BT
4) ???

* 1) Congestion

I know, I'm breaking the habit of naming point 1 "Net status", but
this week "congestion" seems appropriate.  The network itself has
been doing pretty well, but as the bittorrent usage increased, things
started getting more and more clogged up, leading to an essential
congestion collapse [1].

This was expected, and only reinforces our plan - get the new
streaming lib out there, and revamp our tunnel management so we have
sufficient data about peers to use when our fast peers fail.  There
were some other factors in play in the recent network problems, but
the bulk can be traced to the congestion increase and resulting
tunnel failures (which in turn caused all sorts of wild peer
selection).

[1] http://en.wikipedia.org/wiki/Congestion_collapse

* 2) Streaming

There has been a lot of progress with the streaming lib, and I've got
a squid proxy rigged up to it through the live net that I've been
using it frequently for my normal web browsing.  With mule's help,
we've been hammering the streams pretty hard too by piping frost and
FUQID through the network (my god, I never realized how abusive frost
was before doing this!)  A few significant longstanding bugs have
been tracked down this way, and some tweaks to help control massive
numbers of connections have been added.

Bulk streams are working great too, with both slow start and
congestion avoidance, and the quick send/reply connections (ala HTTP
get+response) are doing exactly what they should.

I expect we'll draft some volunteers to try deploying it further over
the next few days, and hopefully get us to the 0.4.2 level soon.  I
don't want to say it'll be so good that it does your dishes, and I'm
sure there'll be bugs that slip through, but it does look promising.

* 3) BT

Barring the recent network troubles, the i2p-bt port has been making
leaps and bounds.  I know a few people have pulled down over a GB of
data through it, and performance has been as expected (due to the old
streaming lib, ~4KBps per peer in the swarm).  I try to listen in on
the work being discussed in the #i2p-bt channel - perhaps duck could
give us a summary in the meeting?

* 4) ???

Thazzit from me for now.  See y'all in the meeting in a few minutes.

=jr

-----BEGIN PGP SIGNATURE-----
Version: PGP 8.1

iQA/AwUBQZpoZBpxS9rYd+OGEQJ7hQCgm635Z/qWpcfDiKQE2JO2Q3eAR/UAn2yQ
ZEawa8wEMLl1tz/uk4BTENkb
=ZS5w
-----END PGP SIGNATURE-----
_______________________________________________
i2p mailing list
i2p at i2p.net
http://i2p.dnsalias.net/mailman/listinfo/i2p

----- End forwarded message -----
--
Eugen* Leitl <a href="http://leitl.org">leitl</a>
______________________________________________________________
ICBM: 48.07078, 11.61144            http://www.leitl.org
8B29F6BE: 099D 78BA 2FD3 B014 B08A  7779 75B0 2443 8B29 F6BE
http://moleculardevices.org         http://nanomachines.net

[demime 1.01d removed an attachment of type application/pgp-signature]