> From: Gianni Johansson <giannijohansson at attbi.com> > To: devl at freenetproject.org > Subject: Re: [freenet-dev] Growing pains -- Better transport level DOS > resistance > Date: 30 Oct 2002 10:33:05 -0500 > > On Wednesday 30 October 2002 07:24, you wrote: > > -----BEGIN PGP SIGNED MESSAGE----- > > > > On Tue, 29 Oct 2002 11:23:07 -0800 Michael Wiktowy <mwiktowy at gmx.net> > > wrote: > > >Never discount the possiblity of sabotage. I hate to sound paranoid > > >but > > >the load on Freenet came on a little too high a little too fast > > >to be a > > > > You may have a point, i was pegged to the wall yesterday with > > connections, and there were quite a few telenet discordian type > > connections on the inbound FNP port and only on that port,, > > > > Sorry i already deleted yesterdays log or i could show you, but > > it was rejection of connection of text of type "sally loved her moose > > fred",,,kinda discordian rambling. > > > > From seednodes and routing tables it'd be easy to get a list of ip's > > and ports to attack. > > Maybe we need a "Turkey trap" filter that keeps track of hosts that > repeatedly make connections that fail with authorization errors and blocks > them at the transport level.
I would say that would be prudent. It might not have to block them permenently (it could just ignore them for a while) just in case a node is temporarily misbehaving due to bugs. The turkey trap could also keep track of the number of threads that are generated by a node and limit per IP. In theory, a node should just make one connection to another node and multiplex the communication across the one line. There is no good reason for one node to be responsible for lots of threads on the recipient. I don't know the details of the new load balancing code so maybe that concept is actually incorporated already. Mike _______________________________________________ devl mailing list devl at freenetproject.org http://hawk.freenetproject.org/cgi-bin/mailman/listinfo/devl
