On Mon, Dec 17, 2001 at 06:48:55PM -0500, Scott Young wrote:
> On Monday 17 December 2001 05:13 pm, you wrote:
> > Another question to bore you guys ;)
> >
> > I understand a message with ahops-to-live value of 1 is sometimes still
> > forwarded on, to reduce the information an attacker could get out of it.
> > What I don't understand, is what information could be learned from the
> > fact thatsomeone will be the last node reached on the path.
> >
> > For example, why would I want to know that some node at the end of the
> > path replies with a Reply.NotFound message? What kind of discriminating
> > information could I get from that?
> 
> It is for malicious ADJACENT nodes.  If node A has a direct connection to 
> node B, node A can find out if the data exists on that node by requesting the 
> file with a htl of 1.  This knowlege can be bad, as Sebastian pointed out.  
> With the certain probibility of forwarding HTL 1 requests, the act of 
> requesting the data can put the data on the node.  This has the benefit of 
> mirroring the data and improving legal arguments.
> 
Additionally, site insertors tend to use a fixed HTL; randomly not reducing the
HTL provides substantial extra anonymity for them.
> Actually, why isn't the HTL concept abonded and replaced with a high certain 
> probability?  This way nodes can't choose a HTL that they can use for gaining 
> data from freenet.  Here is a table of probabilities and and average hops:
> 
> 50%           1 hop
> 70%           1.9 hops
> 80%           3.1 hops
> 90%           6.6 hops
> 95%           13.5 hops
> 96%           17.0 hops
> 97%           22.8 hops
> 98%           34.3 hops
> 99%           68.0 hops
> 
> And if you want to know how I got theese values, here's the formula:
> 1.  probability^hops=0.5   (0.5=50%)
> 2.  log(probability^hops)=hops*log(probability)=log(0.5)
> 3.  hops=log(0.5)/log(probability)
> 
> 
> If a node removes the request by a probabilistic chance, it could return with 
> two messages, one saying the probabilty finally told it to stop searching and 
> another to tell that it could not find a closer node.  If the probability did 
> not fail, nodes could then continue the search with their next-closest node.
> 
> 
> Scott Young
> 
> _______________________________________________
> freenet-tech mailing list
> [EMAIL PROTECTED]
> http://lists.freenetproject.org/mailman/listinfo/tech

-- 
The road to Tycho is paved with good intentions

_______________________________________________
freenet-tech mailing list
[EMAIL PROTECTED]
http://lists.freenetproject.org/mailman/listinfo/tech

Reply via email to