[tor-dev] Proposal xyz : Count Unique IP addresses in an anonymous way

[Nick Mathewson]

This looks plausible to me.  As a formatting note: we usually don't
give the non-proposed solutions equal treatment with the proposed
solutions. Instead, we usually mark them as not proposed.  From this
document, I _think_ you're proposing PCSA, and suggesting that we not
do the others?  But I'm not sure.

For the PCSA portion, we should really have citations to the actual
papers where these formulas and algorithms come from.

We should also consider how this proposal would interact with other
proposed secure aggregation solutions, like Privcount [1] and/or some
other kind of PrivEx [2].  I'd like to hear what the designers of
those ideas think of this one.


[1] https://github.com/privcount/privcount
[2] http://cacr.uwaterloo.ca/techreports/2014/cacr2014-08.pdf