From mboxrd@z Thu Jan 1 00:00:00 1970 From: michi1@michaelblizek.twilightparadox.com Subject: [RFC] new connection-oriented layer 3+4 protocol Date: Wed, 19 Sep 2012 19:08:53 +0200 Message-ID: <20120919170853.GJ2228@grml> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii To: netdev@vger.kernel.org, davem@davemloft.net Return-path: Received: from michaelblizek.twilightparadox.com ([193.238.157.55]:50093 "EHLO michaelblizek.twilightparadox.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1756645Ab2ISR3v (ORCPT ); Wed, 19 Sep 2012 13:29:51 -0400 Content-Disposition: inline Sender: netdev-owner@vger.kernel.org List-ID: Hi! I am programming a new layer 3+4 protocol for mesh networks. Basically it is - connection oriented - source (+optional onion) routed - doing windowing between each neighbor on every connection (no congestion avoidance) - using credits as a metric for QoS The protocol especially tries to withstand DoS attacks and rough users/application without human interference. This includes protection against inserting bogus routes, blackholing/throttling, addressing conflicts (node address can be a hashed public crypto key) and ressource contention. In theory it could provide Byzantine robustness, but in practice the overhead for calculating routes and ressource contention will limit how far this gets. Right now the layer 3 is running somehow (without any crypto). The layer 4 is currently rather incomplete and broken. Project home: http://michaelblizek.homelinux.net/projects/cor/index.html Code: http://repo.or.cz/w/cor.git/tree/HEAD:/net/cor -Michi