[p2p-hackers] [i2p] weekly status notes [sep 28] (fwd from jrandom@i2p.net)

Eugen Leitl eugen at leitl.org
Tue Sep 28 18:10:19 UTC 2004

----- Forwarded message from jrandom <jrandom at i2p.net> -----

From: jrandom <jrandom at i2p.net>
Date: Tue, 28 Sep 2004 10:18:32 -0700
To: i2p at i2p.net
Subject: [i2p] weekly status notes [sep 28]

Hash: SHA1

Hi y'all, weekly update time

* Index:
1) New transport
2) 0.4.1 status
3) ???

* 1) New transport

The 0.4.1 release has been taking longer than expected, but the new
transport protocol and implementation is in place with everything
that has been planned - IP detection, low cost connection
establishment, and an easier interface to help debug when
connections are failing.  This is done by completely throwing out
the old transport protocol and implementing a new one, though we've
still got the same buzzwords (2048bit DH + STS, AES256/CBC/PKCS#5).
If you'd like to review the protocol, its in the docs [1].  The new
implementation is also a lot cleaner, since the old version was just
a bunch of updates accumulated over the last year.


Anyway, there are some things in the new IP detection code that are
worth mentioning.  Most importantly, it is entirely optional - if
you specify an IP address on the config page (or in the
router.config itself), it will always use that address, no matter
what.  However, if you leave that blank, your router will let the
first peer it contacts tell it what its IP address is, which it will
then start listening on (after adding that to its own RouterInfo and
placing that in the network database).  Well, thats not quite true -
if you haven't explicitly set an IP address, it will trust anyone to
tell it what IP address it can be reached at whenever the peer has
no connections.  So, if your internet connection restarts, perhaps
giving you a new DHCP address, your router will trust the first peer
it is able to reach.

Yes, this means no more dyndns.  You're still of course welcome to
keep using it, but its not necessary.

However, this does not do all that you want - if you have a NAT or
firewall, knowing your external IP address is only half of the
battle - you still need to poke the hole for the inbound port. But,
its a start.

(as an aside, for people running their own private I2P networks or
simulators, there is a new pair of flags to be set
i2np.tcp.allowLocal and i2np.tcp.tagFile)

* 2) 0.4.1 status

Beyond the items on the roadmap for 0.4.1, I want to get a few more
things in there - both bugfixes and network monitoring updates.  I'm
tracking down some excessive memory churn issues at the moment, and
I want to explore some hypotheses about the occational reliability
issues on the net, but we'll be ready to roll out the release soon,
perhaps thursday.  It unfortunately will not be backwards compatible,
so it'll be a little bumpy, but with the new upgrade process and the
more forgiving transport implementation, it shouldn't be as bad as
the previous backwards incompatible updates.

* 3) ???

Yeah, we've had short updates the last two weeks, but thats because
we're in the trenches focusing on the implementation, rather than
various higher level designs.  I could tell you about the profiling
data, or the 10,000 connection tag cache for the new transport, but
thats not so interesting.  However y'all may have some additional
things to discuss, so swing on by the meeting tonight and let 'er


Version: PGP 8.1


i2p mailing list
i2p at i2p.net

----- 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
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 198 bytes
Desc: not available
Url : http://zgp.org/pipermail/p2p-hackers/attachments/20040928/d1d9ccee/attachment.pgp

More information about the P2p-hackers mailing list