Re: [Secdispatch] Deterministic generation of public key pairs

[Phillip Hallam-Baker <[e-mail]>]

On Tue, Mar 10, 2020 at 1:28 PM Michael Richardson <[e-mail]>
wrote:

>
> Phillip Hallam-Baker <[e-mail]> wrote:
>     > udf config ssh-agent ZAAA-FFQA-3LE5-SAHG-E6K6-HOTN-TVLB-K4A
>
>     > This is a seed that can be written down and can be used to generate a
>     > private keypair using any of the commonly used public key
> algorithms. So
>     > you can use it for any application where you would use traditional
> key
>     > escrow.
>
>     > One thing I use this for is to move S/MIME and PGP keys about. Can
> also use
>     > it to keep a paper back up of whole disk encryption keys etc.
>
>     > The spec (but not the tool yet) also supports Shamir Secret Sharing.
> I need
>     > to rejig that to match the developments in the threshold spec.
>
>
>     > So is this something we should do and if so where? This is separable
> from
>     > the Mesh and does have some functionality overlap. But it also makes
> a lot
>     > of common sysadmin config much easier.
>
> I think AD sponsorship for the entire UDF document.
> (Not just the deterministic key generation part. or was that part of your
> goal?)
>

I would like to do the whole doc can drop Secure Internet Names which are a
bit on the edge. The digest function parts could be bikeshedded of course
but probably not usefully.



> A better name is needed, as there has been snakeoil created that had a
> similar name.
>

I am OK with the IETF calling it anything people like except for Fred.

If isn't really a fingerprint format at this point either.