Re: An idea about Java and remailer clients and servers...

Phil Fraering writes:
>  Why not "charge" for the ability to send an anonymous message with
>  the duty to have for a short time (maybe an hour or two) running
>  on your machine a node in a remailer network?

User X on Machine A sends a form via HTTP (or a variant- SHTTP, HTTPS, etc.)
to Machine B.  User Y on Machine C receives an anonymous mail from Machine
B.  Suspecting User X, User Y sends a mail to be anonymized and sent back to
himself to User X.  User X's temporary remailer does as it's told.  User Y
now has a strong reason to suspect User X has sent the said mail.

Cpunks write code and all, but I don't think this one's going to work. :-(

> 
> Phil Fraering writes:
> >  Why not "charge" for the ability to send an anonymous message with
> >  the duty to have for a short time (maybe an hour or two) running
> >  on your machine a node in a remailer network?
> 
> User X on Machine A sends a form via HTTP (or a variant- SHTTP, HTTPS, etc.)
> to Machine B.  User Y on Machine C receives an anonymous mail from Machine
> B.  Suspecting User X, User Y sends a mail to be anonymized and sent back to
> himself to User X.  User X's temporary remailer does as it's told.  User Y
> now has a strong reason to suspect User X has sent the said mail.

  If the "duty" cycle is 1 hour and there are 10000 users utilizing
the network, that tells you nothing. All it does it confirm that
User X sent a remailer message within the last hour. One could just
as easily finger User X and use the same reasoning.

  And if one has to suspect User X in the first place, User X has already
blown his cover partially (either by writing style or other leaks)

-Ray

> 
> Phil Fraering writes:
> >  Why not "charge" for the ability to send an anonymous message with
> >  the duty to have for a short time (maybe an hour or two) running
> >  on your machine a node in a remailer network?
> 
> User X on Machine A sends a form via HTTP (or a variant- SHTTP, HTTPS, etc.)
> to Machine B.  User Y on Machine C receives an anonymous mail from Machine
> B.  Suspecting User X, User Y sends a mail to be anonymized and sent back to
> himself to User X.  User X's temporary remailer does as it's told.  User Y

 I forgot to add. There is no reason User X has to run his remailer
immediately. His software could simply commit to running a remailer for
1 hour at some specified future date < some threshold. Any messages
sent to him for remailing would be queued until that time. Therefore,
all your technique would tell you is that the user remailed a message
sometime between date X and date Y. if Y-X > few days to week or two,
the intelligence gathered on User X is miniscule. Traffic analysis would
detect User X using the remailer network anyway.

-Ray

Anonymous writes, concerning the "you want a remailer, you run one" idea:

   User X on Machine A sends a form via HTTP (or a variant- SHTTP, HTTPS, etc.)
   to Machine B.  User Y on Machine C receives an anonymous mail from Machine
   B.  Suspecting User X, User Y sends a mail to be anonymized and sent back to
   himself to User X.  User X's temporary remailer does as it's told.  User Y
   now has a strong reason to suspect User X has sent the said mail.

   Cpunks write code and all, but I don't think this one's going to work. :-(

I was thinking in terms of User X running one node in a mixmaster
network. AFAIK, mixmaster doesn't work in one-bounce mode; otherwise,
why would it go through all the trouble of breaking up the messages,
etc.?

Phil

Phil Fraering writes:
> I was thinking in terms of User X running one node in a mixmaster
> network. AFAIK, mixmaster doesn't work in one-bounce mode; 

The sender can ask the client to set the number of hops as low as 1, if 
she/it/he so desires.

I think Doug Barnes has suggested the best protocol for handling "co-op
remailing" ;)

-Futplex <futplex@pseudonym.com>