[Openid-specs-heart] HEART Profiles - OpenID Connect Certification
Justin Richer
jricher at mit.edu
Wed Dec 2 00:42:04 UTC 2015
I agree with John. There are things that are required to be disabled in HEART that are required to be present in the baseline Connect, and probably vice versa.
But ultimately it’s much too early to fret about that part of the conversation. We’ll burn that bridge when we come to it.
— Justin
> On Dec 1, 2015, at 5:53 PM, John Bradley <ve7jtb at ve7jtb.com> wrote:
>
> I think HEART implementations will be able to pass a profile of the connect certification.
>
> Some things like being required to support http basic authentication, may not be appropriate to test fro HEART deployments if they MUST have that turned off and only use asymmetric
> authentication as an example.
>
> So it may be that a compliant HEART provider will not pass the Connect Dynamic profile.
>
> HEART hasn’t made any decisions about conformance or deployment testing, so it is a bit premature to speculate.
>
> However if I must it will probably require a new profile, much like the GSMA has created for Mobile Connect. That one is something that we need to keep from diverging.
>
> John B.
>
>> On Dec 1, 2015, at 6:15 PM, Mike Jones <Michael.Jones at microsoft.com <mailto:Michael.Jones at microsoft.com>> wrote:
>>
>> I think it would be a sign of a significant problem if implementations of the HEART OpenID Connect profile can’t pass the pertinent OpenID Connect certifications. If people think that that is going to be the case, we should call out the reasons why as early as possible and jointly review them in the HEART and Connect working groups and see what can be done to remedy the situation. Remedies could involve spec work, certification work, or both.
>>
>> I don’t think that any of us want to have to have to explain to people why implementations of the HEART OpenID Connect profile can’t pass the pertinent OpenID Connect certifications. Code reuse is fine as far as it goes, but if we have to go there, it’s a sign of a significant failure on the part of the OpenID Foundation as a whole, at least as I see it.
>>
>> Hopefully there will be no impediments and all of this will have been an academic discussion, but if there are blocking issues, we need to get ahead of it as soon as possible – before the HEART specs go to Implementer’s Draft review.
>>
>> Best wishes,
>> -- Mike
>> <>
>> From: Justin Richer [mailto:jricher at mit.edu <mailto:jricher at mit.edu>]
>> Sent: Tuesday, December 1, 2015 12:25 PM
>> To: Mike Jones <Michael.Jones at microsoft.com <mailto:Michael.Jones at microsoft.com>>
>> Cc: Adrian Gropper <agropper at healthurl.com <mailto:agropper at healthurl.com>>; openid-specs-heart at lists.openid.net <mailto:openid-specs-heart at lists.openid.net>
>> Subject: Re: [Openid-specs-heart] HEART Profiles - OpenID Connect Certification
>>
>> I actually think the conformance will be separate even in the OIDC case, but that’s not something that this group has decided yet and so it’s not something we can say for sure one way or the other.
>>
>> However, that doesn’t imply that the conformance tests or tools will be rewritten from scratch. Code re-use is still very powerful without formal cross-reference.
>>
>> — Justin
>>
>> On Dec 1, 2015, at 3:20 PM, Mike Jones <Michael.Jones at microsoft.com <mailto:Michael.Jones at microsoft.com>> wrote:
>>
>> I assume that what you’re actually saying, Justin, is that OpenID Connect certification wouldn’t be appropriate for the OAuth or UMA profiles but that at least some Connect conformance profiles would be applicable for the OpenID Connect HEART profile. I say that, because I doubt that HEART would want reproduce all the tests verifying standard OpenID Connect conformance. Is that correct or is there something I’m not understanding?
>>
>> Thanks,
>> -- Mike
>>
>> From: Openid-specs-heart [mailto:openid-specs-heart-bounces at lists.openid.net <mailto:openid-specs-heart-bounces at lists.openid.net>] On Behalf Of Justin Richer
>> Sent: Tuesday, December 1, 2015 9:26 AM
>> To: Adrian Gropper <agropper at healthurl.com <mailto:agropper at healthurl.com>>
>> Cc: openid-specs-heart at lists.openid.net <mailto:openid-specs-heart at lists.openid.net>
>> Subject: Re: [Openid-specs-heart] HEART Profiles - OpenID Connect Certification
>>
>> On Dec 1, 2015, at 12:11 PM, Adrian Gropper <agropper at healthurl.com <mailto:agropper at healthurl.com>> wrote:
>>
>> This is a subthread specific to the OIDC Certification issues in the 3 profiles currently up for discussion.
>>
>> OIDC certification has nothing to do with HEART compliance, except that possible future HEART certification systems will follow that model. We’re not requiring compliance with OIDC Certification, especially not for the Oauth2 profile.
>>
>>
>>
>>
>> I'm trying to understand the HEART profile for OAuth 2.0 has numerous mentions of OpenID Connect including:
>> "The authorization server MUST provide an OpenID Connect service discovery endpoint listing the components relevant to the OAuth protocol:"
>> as it relates to real-world implementations of an authorization server in the context of the HEART Use Cases.
>>
>> Again, I ask you to point out the “numerous mentions” and why those are problems as I’ve already explained why we’re using the discovery service from OIDC.
>>
>>
>>
>>
>> The OpenID Certification page http://openid.net/certification/ <http://openid.net/certification/> lists both Google and MITREid Connect. The key difference seems to be that OP Dynamic is not implemented by Google.
>>
>> In the context of building a resource owner's authorization server like HIE of One, the AS wants to make it easy and clear as it decides to add trusted OPs to its OP whitelist.
>>
>> Adding Google as an OP is certainly fussy. The steps involve access by the RO to a Credentials Page on their OP as detailed https://developers.google.com/identity/protocols/OpenIDConnect?hl=en <https://developers.google.com/identity/protocols/OpenIDConnect?hl=en> This is hardly a good user experience for a consumer that simply want to tell her authorization server to trust Google as a source of user authentication.
>>
>> Agreed, but that’s something to bug Google about. It also means that you’ll be shipping your client credentials around with each copy of “HIE Of One” if you don’t want the individual copies to re-register by hand.
>>
>> Furthermore, Google doesn’t support the private_key_jwt method required by HEART at the moment anyway, so it’s a bit of a moot point.
>>
>>
>>
>>
>> I presume that OPs that implement OP Dynamic such as MITREid Connect improve the fussy Google user experience. Let's consider a hospital called NPE (the resource server) that is willing to act as source of user authentication for access to their HEART-compliant API.
>> 1 - Alice (RO) would start by logging in to the NPE (RS) patient portal
>> 2 - Alice would provide the RS with something like a URI or an email address that enables the HEART-compliant RS to discover Alice's HEART-compliant AS (HIE of One).
>> 3 - Alice's AS would put up some kind of authorization form listing NPE as a willing OP Dynamic identity provider for any provider that NPE is willing to take responsibility for authenticating.
>> 4 - If Alice approves, this authorization form, then NPE is added to her AS whitelist of OPs.
>>
>>
>> That would work, but I don’t think it’s a full information flow. I know you’re not trying to show details here but I think that’s going to be required for this proposed system to be real. Best way to do that? Build it and run it! Also, it doesn’t mean that it’s whitelisted, it just means that it’s usable after being discovered and registered. This can all be done alongside statically registered systems, too. A whitelist means that users aren’t prompted for decisions, but if someone else on Alice’s OP logs in, you’d want to prompt them.
>>
>>
>>
>> If we're all together this far, we come up with some clarifying questions:
>>
>> A - Why doesn't any well-known name on the OpenID Certified list implement OP Dynamic?
>>
>> They have a belief that they don’t have to: all the RPs will come to them and they can control the dynamic. There are spurious justifications for this at most providers.
>>
>>
>>
>>
>> B - If HIE of One could get the Django help we need to implement OP Dynamic would the sequence 1-4 above be testable against healthauth.org <http://healthauth.org/> with (alice/wonderland)?
>>
>> Probably.
>>
>>
>>
>>
>> C - When RSs implement the HEART profiles as currently proposed, will it be possible for Alice's AS to combine the authorization for NPE OP registration and NPE resource registration into a single form such as:
>>
>> <image.png>
>> ?
>>
>> The short answer is yes. You’re conflating the form and the functionality that it provides. One form can trigger many things with the right server.
>>
>> — Justin
>>
>>
>>
>>
>> Adrian
>>
>>
>>
>>
>>
>>
>> --
>>
>> Adrian Gropper MD
>>
>> PROTECT YOUR FUTURE - RESTORE Health Privacy!
>> HELP us fight for the right to control personal health data.
>> DONATE: http://patientprivacyrights.org/donate-2/ <http://patientprivacyrights.org/donate-2/>
>> _______________________________________________
>> Openid-specs-heart mailing list
>> Openid-specs-heart at lists.openid.net <mailto:Openid-specs-heart at lists.openid.net>
>> http://lists.openid.net/mailman/listinfo/openid-specs-heart <http://lists.openid.net/mailman/listinfo/openid-specs-heart>
>>
>>
>> _______________________________________________
>> Openid-specs-heart mailing list
>> Openid-specs-heart at lists.openid.net <mailto:Openid-specs-heart at lists.openid.net>
>> http://lists.openid.net/mailman/listinfo/openid-specs-heart <http://lists.openid.net/mailman/listinfo/openid-specs-heart>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openid.net/pipermail/openid-specs-heart/attachments/20151201/224330c7/attachment-0001.html>
More information about the Openid-specs-heart
mailing list