Re: Batching Strategy and Network Structure

[Roger Dingledine <arma@mit.edu>]

I'm going to go to sleep for ~8 hours now, folks.
I leave the paper in your hands.
And we'll still have ~8 hours after that before I submit the final
version.

On Tue, May 07, 2002 at 05:54:11AM +0000, David Hopwood wrote:
> > This is a very tempting idea. It resolves a lot of the
> > diverse-user-behavior problems.
> 
> Exactly - it makes it easy to enforce that all information used to decide
> how the route is chosen is up-to-date. For example:

I agree with a lot of the stuff in this mail. It isn't to the point
where I want to call it part of the Mixminion design, but it's certainly
a lot of stuff that needs to be put in front of people so they think
about it. We have space for a summary of issues and problems and some
approaches to solutions. Section 7 is a good place for that.

Very good stuff. Just make sure to put an "it seems" here and there
if you write it in the paper. Andreas is a reviewer. ;) And we should
start looking at writing this into a second paper. It's really really
important material.

> The optimal structure depends a lot on how the available bandwidth
> constrains the actual width. (I did some back-of-the-envelope
> calculations that suggested even a single cascade could handle the
> current traffic through cypherpunks and Mixmaster, but I can't find
> those notes, and I'd better post this now.)

Web-mixes handle a whole lot of web traffic; they're a deployed cascade.

Thanks,
--Roger