On Wed, Dec 12, 2001 at 01:10:34PM +0100, Nico Galoppo wrote:
<snip>
> This implicits public-key encryption along the mixed path up until it
> reaches the endpoint of the pre-routing phase, optionally requiring a
> pki type key distribution. Have the performance implications of this
> approach been studied before. One could think that it would severely
> downgrade network performance, as assymmetric encryption of large blocks
> of data takes alot of time. Is this kind of approach feasible in a
> real-time application such as Freenet?
i'm running a project (http://www.authnet.org/anonnet/) that implements
a pipenet (a type of mix-net). crypto primitives these days
are pretty fast. AES, for example, is much faster than DES.
the bottle-neck w/ these types of things is latency,
and for strong anonymity you have bandwidth concerns w/ regards
to padding and the like.

<snip>
> Local eavesdroppers and servers are unable to learn the true source of a
> request because it is equally likely to have originated from any member
> of the crowd, and indeed collaborating crowd members cannot distinguish
> the originator of a request from a member who is merely forwarding the
> request on behalf of another. Ofcourse, every message on the path is
> encrypted to prevent local eavesdroppers along the path learning the
> intended receiver of a request.
i was not aware that crowds encrypted anything. if you mean
node-to-node, than one bad node exposes *all* its traffic. add
a couple of bad nodes working together... then encryption is almost
worthless. the problem w/ crowds is it is too easy to trace
traffic, period. anybody who can trace freenet requests
could just as easily trace crowds traffic.

> Moreover, initiators in a Crowds network do not require network topology
> knowledge, while mixmaster initiators do. 
knowing your topology needn't mean *everything*. i'm not sure
what type of mix-net your referring to. are you referring to some trust
protocols? in any event, a freenet or gnutella style discovery mechanism
may suffice, and any vulnerabilities are the lesser of crowds'.

 
Bill

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

Reply via email to