On Thu, Jan 24, 2002 at 12:10:14PM -0600, Edgar Friendly wrote:
> > > > Adding a reason code to the QRej is probably a good idea on general 
> > > > principle.
> > > > 
> > > For the purposes of debugging, it's fine, but I like the fact that now
> > > you can't use QRej responses to determine whether a node is part of a
> > > request chain.  This level of uncertainty I find really useful in
> > > protecting anonymity.
> > 
> > I don't follow your claim.  If a node replies "QRej: route not found"
> > you know it hasn't seen the request before, but if it replies
> > "QRej: looped request" you know it has.
> > 
> > So how does adding a reason code to QueryRejected compromise anonymity?
> > -tc
> > 
> My argument was that now you just get back a generic "QRej", without a
> reason, you don't know which is the case.  So an attacker can't use that
> information to determine the path of a request.
> 
> It doesn't "compromise anonymity", it just gives an attacker more
> information than I'd like them to have.

Well, they already get a reason string that says "route not found" or
"looped request," and they already get the HTL value at the rejecting
node.

-tc


_______________________________________________
Devl mailing list
Devl at freenetproject.org
http://lists.freenetproject.org/mailman/listinfo/devl

Reply via email to