Data Haven problems

[Paul J. Ste. Marie]

At 07:25 PM 1/9/95, dfloyd at io.com wrote:
> ... Of course, the DH will be hidden by a good remailer (anon.penet.fi), but
>it is trivial to use traffic analysis to find where the DH lies.  Just
>monitor traffic from/to the remailer and do a series of store/retrives.
>Then for confirmation, forge a mail from the dh site to the remailer with
>the password (obtained from sniffing) to yourself. ...

Hmm, hmm.  Using c'punk remailers with encrypted send blocks fixes one 
problem, especially if the c'punk mailers do some sort of file splitting and 
reassembly along the lines of what happens to IP packets that are too large 
for a given link.  What would also help would be a mechanism for randomly 
varying the encrypted send-to block.  The password replay attacks can be 
fixed by encrypting the transmitted password along with a timestamp/sequence 
number.

One problem that remains would be a trail left by the increased traffic 
to/from a DH vs a normal user.  That could only be fixed by a multitude of 
DH sites.

    --Paul J. Ste. Marie
      pstemari at well.sf.ca.us, pstemari at erinet.com