On Tue, Oct 28, 2003 at 01:15:04PM +0100, Some Guy wrote:
>  --- Toad <[EMAIL PROTECTED]> wrote: 
> > On Mon, Oct 27, 2003 at 01:18:14PM +0100, Some Guy wrote:
> > >  --- Martin Stone Davis <[EMAIL PROTECTED]> wrote: 
> > > > In any case, I think this thread is complete... Toad's premix routing 
> > > > solution (when he's able to get to it) will solve the anonymity problems 
> > > > I raised.
> > > 
> > > Don't want to reil anyone up but......
> > > 
> > > Long term if we implement something like premix routing, we could change the 
> > > underlieing
> > network
> > > to some more efficient DHT as long as it was secure against DSA and other 
> > > censorship attacks
> > > dirrected at the data.
> > 
> > Yeah, right. It would be as vulnerable to attack as are all DHT based
> > systems. It would not adapt to changing usage patterns or losing nodes,
> > it would reduce anonymity and it would generally be a bad thing. And it
> > is going in over my, and ian's, afaict, dead bodies.
> 
> I thought I made it clear in my reply to Ian that I don't mean the academic DHTs.  I 
> just mean we
> can do things to improve resistance to DoS and cancer nodes and performance, without 
> having to
> worry about objections of "wait that screws up anonymity!!!".

Yeah, it changes the game slightly. Just slightly afaics, but maybe more
than that. I doubt we'll see any radical changes to routing/caching
though.
> 
> > > In the end we'd probably wind up with a two layer network dividing and 
> > > conquering the problem.
> > 
> > > One layer would take care of requestor/poster annonymity, and the other would 
> > > take care of
> > safe
> > > storage.
> > > 
> > > I see two nice ways they could interact:
> > > 1) If I build several annoymous tunnels I can send data to the most appropriate 
> > > end node to
> > > improve routing costs some.
> > > 2) The DHT's traffic and topology should provide great cover against several 
> > > atacks on the
> > > anonoymizing layer.
> > > 
> > > I know this is all very long term, but if we've resigned ourselves to implement 
> > > some premix
> > > solution, it seems logical to keep this idea in the back of our heads.
> > 
> > No. Premix routing would be used exclusively on client requests. Not for
> > other requests. Because we don't want to slow down the system even more.
> 
> Right, I was never suggesting the DHT-like network use the premix routing.  I am 
> suggesting they
> use the same topology/connections because that will create tons of cover traffic.

Ok. That's the intention. Interpretting "DHT-like" loosely.
> 
> > And if you want to implement a general purpose TCP anonymization system,
> > go right ahead. But it'll be very difficult. However, you could
> > certainly run a DHT on top of it.
> 
> That link I have you a while back:
> http://www.pdos.lcs.mit.edu/tarzan/overview.html
> If you read thier latest paper, it's interesting how related the two problems are.  
> They have too
> have to create a topology that keeps cancer nodes from encircling a target, so they 
> have to make
> sure nodes are randomly distributed throughout their net.  Sound familiar?
> 
> If they find nice ways to create random graphs for premix routing, we can use those 
> channels for
> freenet routing, and the two would both benifit.

Hmm. Interesting.

-- 
Matthew J Toseland - [EMAIL PROTECTED]
Freenet Project Official Codemonkey - http://freenetproject.org/
ICTHUS - Nothing is impossible. Our Boss says so.

Attachment: signature.asc
Description: Digital signature

_______________________________________________
Devl mailing list
[EMAIL PROTECTED]
http://dodo.freenetproject.org/cgi-bin/mailman/listinfo/devl

Reply via email to