Re: [minion-cvs] Small corrections covering in particular the first half...

[Roger Dingledine <arma@mit.edu>]

On Fri, Feb 28, 2003 at 04:04:24PM -0500, gd@seul.org wrote:
> -about all the messages addressed to old keys. It turns out that the
> +about all the messages addressed to old keys. The
>  number of IDs a node needs to remember between key rotations is not
> +% This is not very formal:
>  too great a burden.

Nick and I did some computations wherein if you're flat-out flooded,
your replay cache grows by about 2GB/month. (The recent decision to
move key sizes from 1024 to 2048 will greatly reduce max throughput and
improve the above number.)

Should we try to work this value in, or just leave things vague?

>  \item \textbf{Integrated directory servers:} Mixmaster uses several \emph{ad hoc}
>  approaches to distribute information about remailer availability, performance, and
>  keys. But the fact that users and remailers operate with
> -different information introduces \emph{partitioning} attacks.  Mixminion
> +different information introduces \emph{partitioning} attacks
> +\cite{XXXX}.  Mixminion

What's a good thing to cite for partitioning attacks? (We sort of came
up with them independently of whoever we might cite...)

>  \item \textbf{Dummy traffic:} Cottrell briefly mentions dummy messages in
>  \cite{mixmaster-attacks}, but they are not part of the specification
> -\cite{mixmaster-spec}. Mixminion uses a simple dummy policy to
> +\cite{mixmaster-spec}. Mixminion fully supports link padding, dummy
> +traffic and uses a simple dummy policy to
>  improve anonymity.

What does it mean to fully support link padding and dummy traffic? I
guess it means that we've got our own protocol so it's possible, whereas
SMTP-based protocols will find it very hard?

--Roger