Privacy Preserving Measurement
charter-ietf-ppm-01

[Ballot comment]
I think that it would be helpful if the work includes operational deployment considerations, i.e., under what scenarios these protocols can be deployed and what privacy properties are preserved/lost.  Perhaps this is meant to already be covered by "The resulting documents shall clearly describe abuse cases
and remaining attacks which are not prevented or mitigated by the protocol.", but possibly it could be more explicit about covering deployment considerations?

[Ballot comment]
[obligatory remark about milestones]

When we say "the PPM protocols will use cryptographic
algorithms defined by the CFRG", does that just refer to
the encryption algorithms and such (i.e., HPKE), or is it
also intended to encompass the core cryptographic techniques
that enable the privacy-preserving property?  If the latter,
does the CFRG actually have any such work underway or planned
that we would be able to use?

Separately, the wording "defined by" may be worth revisiting,
as there are some cases where the CFRG produces a document that
describes an algorithm previously defined elsewhere, whether to
make it more accessible or to provide some endorsement of it.
Using a different word, like "published", might be preferable.

[Ballot comment]

(1) The public health example in the first paragraph seems out of scope for the IETF.

I can imagine how a public health authority might want to measure exposure using the protocols and algorithms produced by this WG, but that should be their decision.  The details of that use case are outside the general core competency of the IETF and should not be part of the considerations.


(2) "The starting point for PPM WG discussions shall be draft-gpew-priv-ppm."

I don't feel comfortable listing an ID as a starting point because it is not a consensus document.  The WG should decide to use the document as a reference or take a completely different path.  The rest of the charter should stand on its own to provide working parameters.


(3) It would be nice to include milestones.

[Ballot comment]
Some suggestions with the hope to improve the charter.

1st paragraph: there are first two examples (browser & health authority) but later only the browser example is used. Suggest to keep two examples.

Minor MD fixes in the list of mechanisms.

The charter mixes the use of "servers" and "aggregators", suggest to use only one (or did I miss part of the architecture?).

Are the deliverables 'just' one or more protocols ? Should there be some problem statements or requirements ?

Hope this helps

-éric