Re: [jose] I-D Action: draft-ietf-jose-json-web-encryption-09.txt

On Thu, Apr 25, 2013 at 4:40 PM, John Bradley <ve7jtb@ve7jtb.com> wrote:

> For multiple recipients using GCM with a single CEK and IV we can
> logically have 0 AAD segments or 1 AAD segment.
>
> That is just the reality of the situation.
>

I would also note that taking the JOSE overhead out of the AAD segment
frees it up for the application to use.

> One possible solution is to the GCM issue is to have 1 AAD segment
> containing the envelope information for all recipients.
> I understand for people wanting 0 AAD segments that will not be there
> first choice.
>
> The downside of this is that you can't incrementally add recipients, they
> all need to be known upfront to include the info in the AAD.
>
> What I don't know is if there is really any compelling use-case for
> incremental addition of recipients without re-encrypting (changing the IV
> atleast)
>

The XMPP use case is precisely this.  When the sender encrypts the message,
it doesn't know the set of public keys that it's sending it to.

So the only other possible alternative is taking the per-recipient info out
of the integrity check.  But if we're going to do that, we might as well go
all the way and get rid of the integrity check altogether.   Applications
that want to integrity protect information can put it in the AAD segment on
their own.

--Richard

>
> John B.
>  .
>
> On 2013-04-25, at 4:14 PM, Mike Jones <Michael.Jones@microsoft.com> wrote:
>
> Hi Richard,****
>
> Actually, there are four goals in play:****
>
> 1.  GCM****
> 2.  Efficient encoding for multiple recipients****
> 3.  Header integrity protection****
> 4.  Independent protection of each recipient’s headers****
>
> Per my response to Russ, by giving up 4, we can achieve 1, 2, and 3.
> (Credit goes to John Bradley for this solution.)****
>
>                                                             -- Mike****
>
> *From:* Richard Barnes [mailto:rlb@ipv.sx]
> *Sent:* Thursday, April 25, 2013 11:14 AM
> *To:* Mike Jones
> *Cc:* Jim Schaad; jose@ietf.org
> *Subject:* Re: [jose] I-D Action:
> draft-ietf-jose-json-web-encryption-09.txt****
> ** **
> Mike, ****
> ** **
> Your facts are right, but your conclusions are wrong.****
> ** **
> We have three mutually incompatible goals here:****
> 1. GCM****
> 2. Efficient encoding for multiple recipients****
> 3. Header integrity****
> ** **
> We can have any two of these, but not all three.  If we try to do all
> three (JWE-08), then we end up with the vulnerability identified in the
> CFRG thread.  So we need to choose which one to get rid of.****
> ** **
> Getting rid of GCM is clearly not the right answer, as evidenced by the
> reaction in this thread.  There are clear, concrete use reasons to support
> multiple recipients, but not for header integrity.  And header integrity
> can be "polyfilled" with an optional feature, for those who are willing to
> break the multiple recipient case.  Clearly, header integrity is the
> weakest link here.****
> ** **
> JWE-09 is the reductio ad absurdum of header integrity.  Let's do the
> logical thing and stop the absurdity.****
> ** **
> --Richard****
> ** **
> ** **
>
> ** **
> On Thu, Apr 25, 2013 at 2:48 AM, Mike Jones <Michael.Jones@microsoft.com>
> wrote:****
> Jim - I am surprised that you would say that my co-authors Eric Rescorla
> or Joe Hildebrand or the working group would advocate using AES GCM in a
> way that would result in severe security vulnerabilities - in particular,
> allowing attackers to obtain the XOR of the messages to multiple recipients
> encrypted using GCM - a vulnerability identified by the CFRG.
>
> Not stating this in the document would seem to me to be highly
> irresponsible, given the brittleness of GCM in this regard, as identified
> by the CRFG.  As I said to Richard Barnes over dinner last night, while
> unpleasant, and possibly surprising to those who aren't familiar to how GCM
> actually works, as an editor, I viewed including the statement that "AES
> GCM MUST NOT be used when using the JWE JSON Serialization for multiple
> recipients, since this would result in the same Initialization Vector and
> Plaintext values being used for multiple GCM encryptions" as necessary, and
> "truth in advertising".
>
>                                 -- Mike****
>
> -----Original Message-----
> From: jose-bounces@ietf.org [mailto:jose-bounces@ietf.org] On Behalf Of
> Jim Schaad
> Sent: Wednesday, April 24, 2013 9:07 PM
> To: Mike Jones
> Cc: jose@ietf.org
> Subject: Re: [jose] I-D Action: draft-ietf-jose-json-web-encryption-09.txt
>
> Mike,
>
> AES GCM MUST NOT be used when using the JWE JSON Serialization for
>    multiple recipients, since this would result in the same
>    Initialization Vector and Plaintext values being used for multiple
>    GCM encryptions.
>
> I doubt your co-authors would agree with this.
> I doubt the working group with agree with this.
> I know that at least one co-chair does not agree with this I can predict
> that the AD and IESG along with the security directorate would crucify me
> if I allowed this to stand in the document..
>
> Jim
>
>
>
> > -----Original Message-----
> > From: jose-bounces@ietf.org [mailto:jose-bounces@ietf.org] On Behalf
> > Of internet-drafts@ietf.org
> > Sent: Tuesday, April 23, 2013 5:29 PM
> > To: i-d-announce@ietf.org
> > Cc: jose@ietf.org
> > Subject: [jose] I-D Action: draft-ietf-jose-json-web-encryption-09.txt
> >
> >
> > A New Internet-Draft is available from the on-line Internet-Drafts
> directories.
> >  This draft is a work item of the Javascript Object Signing and
> > Encryption Working Group of the IETF.
> >
> >       Title           : JSON Web Encryption (JWE)
> >       Author(s)       : Michael B. Jones
> >                           Eric Rescorla
> >                           Joe Hildebrand
> >       Filename        : draft-ietf-jose-json-web-encryption-09.txt
> >       Pages           : 54
> >       Date            : 2013-04-23
> >
> > Abstract:
> >    JSON Web Encryption (JWE) is a means of representing encrypted
> >    content using JavaScript Object Notation (JSON) data structures.
> >    Cryptographic algorithms and identifiers for use with this
> >    specification are described in the separate JSON Web Algorithms (JWA)
> >    specification.  Related digital signature and MAC capabilities are
> >    described in the separate JSON Web Signature (JWS) specification.
> >
> >
> > The IETF datatracker status page for this draft is:
> > https://datatracker.ietf.org/doc/draft-ietf-jose-json-web-encryption
> >
> > There's also a htmlized version available at:
> > http://tools.ietf.org/html/draft-ietf-jose-json-web-encryption-09
> >
> > A diff from the previous version is available at:
> > http://www.ietf.org/rfcdiff?url2=draft-ietf-jose-json-web-encryption-0
> > 9
> >
> >
> > Internet-Drafts are also available by anonymous FTP at:
> > ftp://ftp.ietf.org/internet-drafts/
> >
> > _______________________________________________
> > jose mailing list
> > jose@ietf.org
> > https://www.ietf.org/mailman/listinfo/jose
>
> _______________________________________________
> jose mailing list
> jose@ietf.org
> https://www.ietf.org/mailman/listinfo/jose
> _______________________________________________
> jose mailing list
> jose@ietf.org
> https://www.ietf.org/mailman/listinfo/jose****
> ** **
> _______________________________________________
> jose mailing list
> jose@ietf.org
> https://www.ietf.org/mailman/listinfo/jose
>
>
>