Re: [dmarc-ietf] ARC Crypto Algorithm Selection

[Hector Santos <[e-mail]>]

In my view, ARC, as an "Experimental Status" (and still in design) 
proposal *should* include support for all valid DKIM STD hashing 
methods, including rsa-hash1, and I agree, should not be the impetus 
for removing sha1 support in DKIM implementations.

Until a DKIM implementation is "ready" to a) consider DCRUP and b) add 
support for ed25119-sha256 (which comes with a considerable technical 
barrier), rsa-sha1 is still going to be available and used by pure 
DKIM STD implementations for the foreseeable future (a long time).  I 
haven't been seeing many messages in my logs showing rsa-ed25119 DKIM 
signings, in today's logs so far, not one.

-- 
HLS


On 10/23/2018 11:24 AM, Kurt Andersen (b) wrote:
> On Tue, Oct 23, 2018 at 3:58 AM Scott Kitterman <[e-mail]
> <mailto:[e-mail]>> wrote:
>
>     Last time I looked at this particular issue, ARC could use any
>     algorithm that
>     DKIM uses.
>
>
> Still correct.
>
>        As I recall, that was once of the stimuli for the DCRUP working
>     group (to avoid having rsa-sha1 be valid for ARC by obsoleting it
>     in DKIM).
>
>
> I don't think that ARC drove the DCRUP work as much as the desire to
> get badness officially obsoleted :-)
>
>     It looks like this discussion has been moved to a new draft,
>     https://tools.ietf.org/html/draft-ietf-dmarc-arc-multi-01
>     (although the
>     reference is wrong,
>     https://tools.ietf.org/html/draft-ietf-dmarc-arc-multi-02
>     is current.
>
>
> Yes that's correct.
>
>     Unfortunately, I don't find any actual guidance on what algorithms
>     are
>     currently used.
>
>
> Any that are valid for DKIM - see first sentence above.
>
>     Secion 6, Phases of Algorithm Evolution, gives some process
>     (which seriously needs revision - I thought we all knew flag days
>     don't work
>     at Internet scale), but no actual guidance.
>
>
> Correct - we've deferred any real attention on this until we got the
> ARC protocol document nailed down. With the request to publish being a
> few hours old now from Barry, this might be an excellent topic to
> pursue next :-)
>
>     DKIM, as updated by the DCRUP work, has two valid crypto algorithms:
>
>     rsa-sha256
>     ed25119-sha256
>
>     One has been obsoleted:
>
>     rsa-sha1
>
>     Which among those is valid for ARC and how do I know?
>
>
> The same ones that are valid for DKIM (as updated by DCRUP). What we
> haven't worked out is how to handle mixed algorithm chains.
>
> --Kurt
>
>
> _______________________________________________
> dmarc mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/dmarc
>