[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[freehaven-dev] request for comments: P2P east submission
I've been sitting on this for a week or two now. I wrote it but never
sent it. Comments appreciated. Feel free to suggest other topics
(ideally with an abstract ;) that we should include in the mail.
Subject: P2P east submission
Tim O'Reilly told me that I should propose several different topics,
so you guys could pick which ones you thought would fit best in the
schedule. Since the conference is so far in the future, I figured this
is a good idea. We can also help out on panels, etc, if you think any
of them are appropriate for us. Let me know if you have any questions;
we're pretty flexible. (I've ordered these with the ones you'll like
Tarzan: a Decentralized Stream-based Anonymizing Network (30-60 mins).
We describe a decentralized network which emphasizes simplicity and
extensibility, and allows anonymous streaming connections between
peers. Tarzan is designed to be a low-level infrastructure which provides
either one-way or two-way streams, where one or both endpoints can be
behind firewalls. We show techniques for using Tarzan as an anonymizing
back-end for both clients and servers in other P2P file sharing systems
including Freenet, Free Haven, and Publius.
The Free Haven reputation system (30-60 mins). We cover the variety of
topics that need to be considered and solved to build a robust reputation
system over a P2P system like Free Haven. These include personalized
(local) reputations, rater credibility, and reputation categories (inferring performance for one context based on performance in others).
We examine some related reputation systems (EBay, Advogato) and lessons
that can be reused in other p2p systems (eg Mojo Nation).
Taxonomy of Anonymity for P2P Systems (30-60 mins). We describe
types of anonymity (including reader, publisher, and server), discuss
levels of anonymity and classes of adversaries, explain linkability and
pseudonymity, and show techniques for layering identity protection onto
current p2p systems. We emphasize real-world considerations rather than