Re: [jose] Keys in the documents

[Mike Jones <[e-mail]>]

-14 now describes which RSA key parameters are there to enable optimizations, and states that their presence is RECOMMENDED, and that if any are present, all must be present (yes, with special language for the case of 3 or more prime factors).

                                                            -- Mike

From: Richard Barnes [mailto:[e-mail]]
Sent: Tuesday, July 16, 2013 4:21 PM
To: Brian Campbell
Cc: Mike Jones; Matt Miller (mamille2); Jim Schaad; [e-mail]<mailto:[e-mail]>
Subject: Re: [jose] Keys in the documents

On Tue, Jul 16, 2013 at 4:40 PM, Brian Campbell <[e-mail]<mailto:[e-mail]>> wrote:
I like this change and think it will make it much more straightforward to consume the examples.
One thing I noticed though, in Section 5.3.2 of JWA "JWK Parameters for RSA Private Keys" [1] it says that all the members (excepting "oth") are required for private keys.
However the example JWK RSA keys in JWE [2] and JWS [3] only have the "d" (Private Exponent) Parameter part of the private portion.
Can we relax/change JWA to say something like "d" is always required and either all of others (with the caveat for "oth") are required to be there together or that they all need to be omitted?
The Private Exponent is all that's functionally needed, right? And the rest are optimizations? I honestly don't know much (okay anything) about CRT vs plain old RSA keys. But it seems like there are cases where it'd be totally reasonable to have just the "d" - and the examples in JWS and JWE seem to make that point.

Yes.  This change should be made.  Technically, only the modulus (n) and private exponent (d) are required.  So the requirement levels for a private key would be:
n, d: MUST
e: SHOULD (so that you can derive the corresponding public key)
p,q,dp,dq,qi: MAY (since these are all optimizations)

--Richard


[1] http://tools.ietf.org/html/draft-ietf-jose-json-web-algorithms-13#section-5.3.2
[2] http://tools.ietf.org/html/draft-ietf-jose-json-web-encryption-13#appendix-A.1.4
[3] http://tools.ietf.org/html/draft-ietf-jose-json-web-signature-13#appendix-A.2.1

On Sun, Jul 14, 2013 at 3:03 PM, Mike Jones <[e-mail]<mailto:[e-mail]>> wrote:
FYI - this was done in the -12 drafts.

                                                            -- Mike

From: Mike Jones [mailto:[e-mail]<mailto:[e-mail]>]
Sent: Friday, June 21, 2013 8:58 AM
To: Matt Miller (mamille2); Richard Barnes

Cc: Jim Schaad; [e-mail]<mailto:[e-mail]>; [e-mail]<mailto:[e-mail]>
Subject: RE: [jose] Keys in the documents

Will do.
________________________________
From: Matt Miller (mamille2)
Sent: 6/21/2013 6:06 AM
To: Richard Barnes
Cc: Jim Schaad; [e-mail]<mailto:[e-mail]>; [e-mail]<mailto:[e-mail]>
Subject: Re: [jose] Keys in the documents

+1

On Jun 20, 2013, at 8:48 PM, Richard Barnes <[e-mail]<mailto:[e-mail]>>
wrote:

> +1
>
> On Thursday, June 20, 2013, Jim Schaad wrote:
>
>> Is there any reason not to provide the public and private keys in the
>> appendixes as JWK objects?  This would make them easier to understand and
>> put them into a format that one expects to be understood by JOSE systems.*
>> ***
>>
>> ** **
>>
>> Jim****
>>
>> ** **
>>

- m&m

Matt Miller < [e-mail]<mailto:[e-mail]> >
Cisco Systems, Inc.

_______________________________________________
jose mailing list
[e-mail]<mailto:[e-mail]>
https://www.ietf.org/mailman/listinfo/jose



On Tue, Jul 16, 2013 at 4:40 PM, Brian Campbell <[e-mail]<mailto:[e-mail]>> wrote:
I like this change and think it will make it much more straightforward to consume the examples.
One thing I noticed though, in Section 5.3.2 of JWA "JWK Parameters for RSA Private Keys" [1] it says that all the members (excepting "oth") are required for private keys.
However the example JWK RSA keys in JWE [2] and JWS [3] only have the "d" (Private Exponent) Parameter part of the private portion.
Can we relax/change JWA to say something like "d" is always required and either all of others (with the caveat for "oth") are required to be there together or that they all need to be omitted?
The Private Exponent is all that's functionally needed, right? And the rest are optimizations? I honestly don't know much (okay anything) about CRT vs plain old RSA keys. But it seems like there are cases where it'd be totally reasonable to have just the "d" - and the examples in JWS and JWE seem to make that point.

[1] http://tools.ietf.org/html/draft-ietf-jose-json-web-algorithms-13#section-5.3.2
[2] http://tools.ietf.org/html/draft-ietf-jose-json-web-encryption-13#appendix-A.1.4
[3] http://tools.ietf.org/html/draft-ietf-jose-json-web-signature-13#appendix-A.2.1

On Sun, Jul 14, 2013 at 3:03 PM, Mike Jones <[e-mail]<mailto:[e-mail]>> wrote:
FYI - this was done in the -12 drafts.

                                                            -- Mike

From: Mike Jones [mailto:[e-mail]<mailto:[e-mail]>]
Sent: Friday, June 21, 2013 8:58 AM
To: Matt Miller (mamille2); Richard Barnes

Cc: Jim Schaad; [e-mail]<mailto:[e-mail]>; [e-mail]<mailto:[e-mail]>
Subject: RE: [jose] Keys in the documents

Will do.
________________________________
From: Matt Miller (mamille2)
Sent: 6/21/2013 6:06 AM
To: Richard Barnes
Cc: Jim Schaad; [e-mail]<mailto:[e-mail]>; [e-mail]<mailto:[e-mail]>
Subject: Re: [jose] Keys in the documents

+1

On Jun 20, 2013, at 8:48 PM, Richard Barnes <[e-mail]<mailto:[e-mail]>>
wrote:

> +1
>
> On Thursday, June 20, 2013, Jim Schaad wrote:
>
>> Is there any reason not to provide the public and private keys in the
>> appendixes as JWK objects?  This would make them easier to understand and
>> put them into a format that one expects to be understood by JOSE systems.*
>> ***
>>
>> ** **
>>
>> Jim****
>>
>> ** **
>>

- m&m

Matt Miller < [e-mail]<mailto:[e-mail]> >
Cisco Systems, Inc.

_______________________________________________
jose mailing list
[e-mail]<mailto:[e-mail]>
https://www.ietf.org/mailman/listinfo/jose