[mnet-devel] Re: Experiences Deploying a Large Scale Emergent Network (fwd)

Eugen Leitl eugen@leitl.org
Tue, 9 Apr 2002 15:45:55 +0200 (CEST)

---------- Forwarded message ----------
Date: Tue, 09 Apr 2002 06:16:05 -0700
From: Zooko <zooko@zooko.com>
To: A. Melon <juicy@melontraffickers.com>
Cc: cryptography@wasabisystems.com, mnet-devel@lists.sourceforge.net,
Subject: [mnet-devel] Re: Experiences Deploying a Large Scale Emergent
Resent-Date: Tue, 9 Apr 2002 15:45:30 +0200 (CEST)
Resent-From: <eleitl@leitl.org>
Resent-To: <eugen@hydrogen.leitl.org>
Resent-Subject: [mnet-devel] Re: Experiences Deploying a Large Scale

[This is in reply to a message that was sent to me and to 
<cryptography@wasabisystems.com> on 2002-03-22.  I do not see it on 
mail-archive.com [1], so if you are interested you might want to view the 
archive at sf.net/projects/mnet [2].]

I'd like to thank A. Melon for criticism of my paper "Experiences Deploying a 
Large-Scale Emergent Network".  I've updated the paper in preparation for its 
inclusion in a printed dead-tree proceedings and attempted to address some of 
A. Melon's criticisms.  In particular, I've tried to be more clear about the 
magnitude of Mojo Nation's failures by adding the typical and maximum number 
of simultaneously connected nodes.

I've also added some observations about "two big mistakes that would be easy 
to correct", something I understood only after chatting with the researchers 
at the Peer-to-Peer Workshop.

I've also attempted to address A. Melon's other criticism: that it isn't clear 
which specific issues are most to blame for the overall poor behavior.  I've 
added statements about my belief that the high node churn rate was largely due 
to the poor data availability and that conversely the poor data availability 
was partially due to the high node churn rate.  I've also added a statement 
that there are a lot of important aspects of the system as a whole which are 
omitted from the scope of the paper.  (Including agnostically-blindable 
digital tokens and many other things.)

I'd like to thank A. Melon and the participants of the Peer-to-Peer workshop 
for feedback.  Most of all I'd like to thank the architects of Mojo Nation: 
Jim McCoy and Doug Barnes.  Mojo Nation was a brave experiment, and I hope 
that we will all benefit from the resulting knowledge.

Here is the URL for the current version of the paper:




[1] http://www.mail-archive.com/cryptography@wasabisystems.com/
[2] http://sourceforge.net/mailarchive/forum.php?thread_id=579361&forum_id=7702

Security and Distributed Systems Engineering

mnet-devel mailing list