Re: [CFRG] Changes planned for the next VDAF draft

[Christopher Patton <[e-mail]>]

Hi list,

This is to let you know that the authors plan to cut the next VDAF draft
(01) before completing the Poplar1 spec. This is to unblock development of
the PPM WG's DAP protocol [1]. We're still aiming to finish Poplar1 before
IETF 114, but it'll appear in the next draft.

With that in mind, we're considering one additional syntax change for 01
[2], which should make using VDAFs in a protocol spec a bit more ergonomic.
See the issue linked to [2] for more details.

Thanks,
Chris P.

[1] https://github.com/ietf-wg-ppm/draft-ietf-ppm-dap
[2] https://github.com/cfrg/draft-irtf-cfrg-vdaf/pull/73

On Mon, May 2, 2022 at 8:07 AM Christopher Patton <[e-mail]>
wrote:

> Hi CFRG,
>
> As you may have noticed, the RG recently adopted
> https://datatracker.ietf.org/doc/draft-irtf-cfrg-vdaf/. In this email I
> wanted to give you a heads up about the work the authors would like to get
> done before IETF 114. Please let us know if you object to any of these
> changes or if there is something else you'd like to see in the next draft.
>
>    - Complete spec for Poplar1 (as promised at IETF 113). This would
>    resolve https://github.com/cfrg/draft-irtf-cfrg-vdaf/issues/22 and
>    https://github.com/cfrg/draft-irtf-cfrg-vdaf/issues/44.
>    - Minor improvement to Prio3 that reduces communication overhead:
>    https://github.com/cfrg/draft-irtf-cfrg-vdaf/issues/39. Note that this
>    scheme still requires security analysis.
>    - Define syntax of DAFs, i.e., VDAFs w/o verifiability:
>    https://github.com/cfrg/draft-irtf-cfrg-vdaf/issues/20. This allows us
>    to avoid round trips over the network during the preparation phase in
>    applications where the clients are trusted to shard their inputs properly.
>    - Remove the public parameter from the syntax:
>    https://github.com/cfrg/draft-irtf-cfrg-vdaf/issues/19. This feature
>    was included in order to make room for VDAFs that might use public key
>    cryptography. So far we have none, and the public parameter has turned out
>    to be a bit of a burden for PPM, which right now is the main application of
>    VDAFs.
>
> Thanks,
> Chris P.
>