Thanks, Mike.  I've sent out a question to get the viewpoint of the current
IESG members in hopes to prevent issues as we move forward.  I'll post back
to the discussion once I get enough input on current preferences in case
anything has changed from experience, etc.


On Fri, Jun 13, 2014 at 4:04 PM, Mike Jones <[email protected]>
wrote:

>  This was considered by the WG as issue #10 -
> http://trac.tools.ietf.org/wg/jose/trac/ticket/10.
>
>
>
> In the OAuth context, I know that draft-ietf-oauth-assertions and
> draft-ietf-oauth-saml2-bearer were sent to the IESG for review in 2012 and
> then sent back to the OAuth working group by the IESG because they felt
> that additional work needed to be done on the drafts so that they would
> *actually* be interoperable – not just that it was possible for
> implementations to interoperate.  Stephen Farrell could perhaps speak more
> to that.
>
>
>
> That clear sentiment from the IESG to the OAuth WG has also informed the
> decisions for the JWT and JOSE specs to keep a small set of MTI algorithms
> so that implementations would actually have a common basis to interoperate.
>
>
>
> IESG feedback on requiring interoperability was referenced by Jim Schaad
> in his text closing the issue:
>
> The IESG has had a discussion on this issue as part of the recent charter
> discussions. The chairs believe that is is clear from that discussions that
> MTI algorithms are going to be required in order for the documents to
> progress. For that reason we are closing this ticket.
>
>
>
>                                                             -- Mike
>
>
>
> *From:* OAuth [mailto:[email protected]] *On Behalf Of *Mike Jones
> *Sent:* Friday, June 13, 2014 12:27 PM
> *To:* Kathleen Moriarty; Hannes Tschofenig
>
> *Cc:* [email protected]
> *Subject:* Re: [OAUTH-WG] JWT review
>
>
>
> In no place is SHA-1 or algorithms using it MTI.  You can see the set of
> MTI algorithms by looking at those marked “Required” in the registries.
>
>
>
> A small set of required algorithms is present, with the choices based on a
> detailed survey of what algorithms are widely deployed, to provide a basis
> for implementations to interoperate.  Recognizing that the set of
> algorithms that will be appropriate to have as required will change over
> time, Sean Turner suggested that we enable future drafts to update the
> Implementation Requirements in the registries, with expert review.  (So for
> instance, an algorithm that might be “Required” today could be marked
> “Deprecated” in the future.)  We adopted Sean’s suggestion a good while ago.
>
>
>
> This is another area that was widely discussed within the JOSE working
> group, and there was never consensus to remove the implementation
> requirements, which have always been present.
>
>
>
>                                                             -- Mike
>
>
>
> *From:* OAuth [mailto:[email protected] <[email protected]>] *On
> Behalf Of *Kathleen Moriarty
> *Sent:* Friday, June 13, 2014 12:14 PM
> *To:* Hannes Tschofenig
> *Cc:* [email protected]
> *Subject:* Re: [OAUTH-WG] JWT review
>
>
>
> Hi Hannes,
>
>
>
> Thank you for going through the various reviews, since the JOSE ones
> should be of interest to Oauth.  I'll respond in-line.
>
>
>
> On Thu, Jun 12, 2014 at 4:27 AM, Hannes Tschofenig <
> [email protected]> wrote:
>
> Hi Kathleen,
>
> on the first item I have a few minor remarks: You wrote:
>
> "
> As I read through the Algorithms (JWA) draft there are some changes that
> will need to be made to avoid problems during the IESG review.  This is
> a pretty big change for the draft, but will help make the review and
> approval faster.  Typically, the lists of algorithms are handled through
> a draft update as opposed to creating an IANA registry.  A good example
> is a recent update of a draft in the IPSECME working group so you can
> see the structure and the precedence for this model.
> "
>
> FYI - this is from the start of a long thread that has been worked out
> already.  I had included a link to the JWA review only for the section on
> the security consideratiosn section as many of the drafts in JOSE, and at
> least one in OAuth start out with the same paragraph that could use some
> updating and correcting.  I wanted to make sure this working group was
> aware since JWT shares that same paragraph.  Mike is working through new
> text and has solicited help from the WG (please respond on the JOSE list).
>
>
> The IANA registry for the algorithm serves a different purpose than a
> document recommending the specific algorithms. The reference to the
> IPSECME document only provides the latter. It is also important to note
> that the JWA not only defines the algorithm tags for the IANA registry
> but also explains how they actually work with the JOSE defined JSON
> structures (which is again a difference to the mentioned IPSECME document).
>
>  The discussion on having a registry versus a draft has been settled.
>  The possibility of an issue came to me through an AD and after discussion,
> it is fine as it is.  There were some considerations that needed to get
> surfaced, so the document can remain as-is.  Sorry for the confusion.  I'll
> file this away for the future reference.
>
>
> Of course, the JWA document does both via the IANA registry and there is
> the question about how these recommendations would then get updated and
> what the consensus process is.
>
> In an mail to the JOSE mailing list I argued against any MTI
> recommendations since JOSE is a baseline technology that will be used in
> a variety of different contexts and it is super likely that the
> algorithm requirements will hugely vary.
>
> I am just thinking about what algorithms I would recommend when using
> the JOSE work in an IoT environment. My recommendations would deviate
> from the currently given recommendations, which are largely impacted by
> the Web community.
>
> Here is the mail I sent to the JOSE list:
> http://www.ietf.org/mail-archive/web/jose/current/msg04032.html
>
> So, my recommendation is to
>
> 1) have no MTI requirements in the JWA spec
> 2) remove the 'JOSE Implementation Requirements' column from the IANA
> registry.
>
>
>
> Interesting.   I do remember having these discussions with Sean and
> Richard (see
> http://www.ietf.org/mail-archive/web/jose/current/msg04060.html).  In
> Jim's opinion, (from:
> http://www.ietf.org/mail-archive/web/jose/current/msg04062.html), his
> view is that even the MTI in JWA can be overridden in the spec.  I wonder
> why you would have an MTI then?
>
>
>
> This closed out the discussion and it would be better to see it on the
> JOSE list than here.  If the point is to get Oauth people who are
> encountering conflicts as a user of JOSE drafts to chime in, that should
> happen on the JOSE list.  I suspect this will be an issue for XMPP as well.
>  They are phasing out SHA-1, so if that's MTI for fingerprints, they may
> still feel like they have to support SHA-1 for that purpose even though
> their work specifies that SHA-2 should be used everywhere.
>
>
>
> Since JWA is getting closer to IESG review, I'll ask other ADs their
> thoughts on how they like to see this sort of thing handled.  Both Richard
> and Jim raised valid points.
>
>
>
> Thank you,
>
> Kathleen
>
>
> Ciao
> Hannes
>
>
>
> On 06/09/2014 06:17 PM, Kathleen Moriarty wrote:
> > Hello,
> >
> > I am in process of working through the JOSE drafts and also read the
> > Oauth JWT draft last week.  There is some overlap in text that may
> > require some joint work to correct.
> >
> > 1. For JWT, the Security Considerations section starts off with the same
> > text that is in several of the JOSE drafts.  In my review of the JWA
> > draft, I asked for some fixes that will need to be made to this draft as
> > well.  Here is a link to that review and it may be easier to help with
> > this work in one spot where text will be reused.  Mike has asked the
> > JOSE WG to assist, but it make make sense for Oauth folks to help as
> > well.  If it makes sense, a pointer to existing text is also fine.
> >
> > http://www.ietf.org/mail-archive/web/jose/current/msg04064.html
> >
> > 2. Sections 5.1 and 5.2 are a little confusing.  However, the use of
> > "typ" and "cty" appear in 3 drafts (at least), so this should get
> > addressed with an approach that considers the joint text to reduce
> > confusion for developers.  The initial descriptions are in the JOSE JWS
> > draft, so that may need most of the work, but it also appears in this
> > draft and the JOSE JWK draft.  In my writeup for the JWK review, I
> > listed out some questions and would like to see improvements across
> > these drafts.  This will likely require some joint work and may be best
> > in response to the JWK review to keep it in one place.
> >
> > http://www.ietf.org/mail-archive/web/jose/current/msg04172.html
> >
> > Thank you!
> >
> > --
> >
> > Best regards,
> > Kathleen
> >
> >
>
> > _______________________________________________
> > OAuth mailing list
> > [email protected]
> > https://www.ietf.org/mailman/listinfo/oauth
> >
>
>
>
>
>
> --
>
>
>
> Best regards,
>
> Kathleen
>



-- 

Best regards,
Kathleen
_______________________________________________
OAuth mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/oauth

Reply via email to