[Openid-specs-heart] HEART Scopes & Resource Sets

Tue Jun 16 17:01:22 UTC 2015

The ability in OAuth for a client to request multiple resources and
permissions over them seems valuable, of course. Given that "resource set
registration" and "permission granting" are distinct activities in UMA,
though, I'm reluctant to conflate them for OAuth historical reasons, and
would prefer to think about how to solve the problem "correctly" if
possible. (Josh, btw, I did give examples of distinct scopes over different
resources, e.g., Salesforce's "chatter" scope. This pattern seems pretty
common.)

Though it didn't make it into UMA V1.0, the group discussed a pattern where
the RS could register multiple permissions vs. just a single one, and there
is a communications channel between the client and AS (the request for
authorization data at the RPT endpoint) that could be used to convey the
client's desires as well. I want to be clear that this is in the realm of
potential *UMA extensions* (e.g., for consideration in V.next) rather than
just HEART profiling. But the ease of experimenting with and creating
extensions (e.g. adding JSON properties to requests and responses) is one
big reason why we felt comfortable wrapping up V1.0 after a fairly lengthy
development process.

In fact, FWIW, I've been having a discussion with some folks in a different
context that looks exactly like "generous resource set/permission
registration by the RS at the AS in response to a limited initial access
request by a client". So this isn't a totally one-off conversation.

*Eve Maler*ForgeRock Office of the CTO | VP Innovation & Emerging Technology
Cell +1 425.345.6756 | Skype: xmlgrrl | Twitter: @xmlgrrl
Join our ForgeRock.org OpenUMA <http://forgerock.org/openuma/> community!

On Tue, Jun 16, 2015 at 7:42 AM, Justin Richer <jricher at mit.edu> wrote:

> Eve,
>
> The main difference is that it’s not at all uncommon in the OAuth world to
> ask for authorization to multiple resources protected by the same AS
> simultaneously. In fact, this is seen as a *feature* of the OAuth approach,
> since it’s lower decision overhead for the user (when done right). In that
> case, if a client asks for “read write delete” scopes of an AS, the AS
> still needs to know *what* those scopes apply to. Since OAuth doesn’t have
> any type of resource or audience identifier (a big hole in the spec), this
> gap has been usually filled by having a scope identify the resource. Note
> that this is still semantically sensible and doesn’t go against what
> “scope” is defined as.
>
> This is where you get the matrix definition. You’ve got some scopes that
> mean “where can I do things” and others that mean “what can I do there”. I
> think Josh’s approach of “what.where” is reasonable given this technical
> constraint, and not without precedent. As far as the AS is concerned, it’s
> dealing with just strings from the client, but it can still easily make the
> UX of the authorization page a little smart based on the understood
> semantics of these well-defined scopes.
>
>  — Justin
>
> On Jun 15, 2015, at 7:44 PM, Eve Maler <eve.maler at forgerock.com> wrote:
>
> Hi Josh-- Below...
>
>
> *Eve Maler*ForgeRock Office of the CTO | VP Innovation & Emerging
> Technology
> Cell +1 425.345.6756 | Skype: xmlgrrl | Twitter: @xmlgrrl
> Join our ForgeRock.org OpenUMA <http://forgerock.org/openuma/> community!
>
> On Mon, Jun 15, 2015 at 2:24 PM, Josh Mandel <
> Joshua.Mandel at childrens.harvard.edu> wrote:
>
>> Hi all,
>>
>> I didn't mean to take a hard-line position on today's call about scope
>> definitions! To my mind, our approach to scopes will need to work
>> hand-in-hand with our approach to endpoint (or resource set) discovery --
>> so I feel a bit awkward discussing scopes here in isolation. But that said,
>> let me see if I can at least highlight the tension that we heard in the
>> past hour's discussion (in a neutral way):
>>
>> ---
>> *Goal: Whatever the model, we want to support a use case where Alice
>> signs into her resource server and can set some policies in an intuitive
>> way. |She'd see something like (very, very roughly):*
>>
>>  My Medications:
>>  * Who can view?
>>  * Who can write new prescriptions?
>>
>> My Step Counts
>>  * Who can view?
>>  * Who can remove?
>> ---
>>
>> The question is about how this works under the hood.  I think we were
>> discussing two models:
>>
>> *Model 1: The "UMA-First" approach*
>> *We have a resource set like "Alice's Medications", with scopes like
>> "view" and "prescribe". And we'd have a resource set like "Alice's Step
>> Counts" with scopes like "view" and "delete".*
>>
>> *Model 2: The "OAuth-First" approach*
>> *We have a resource set like "Alice's FHIR Endpoint", with scopes like
>> "Medications.view", "Medications.prescribe", "Steps.view", and
>> "Steps.delete".*
>>
>>
> Talking about an "OAuth-first" approach for setting policies is making me
> confused. I know what it looks like to enable OAuth-like flows in UMA when
> Alice is both the requesting party and the owner of the resource. And I
> know what it looks like to enable Alice to set policies at an UMA
> authorization server (which might hold the results of a previous OAuth-like
> flow done in UMA). But I don't know what "setting policies in OAuth" means
> because the OAuth experience is about consenting at run time (possibly
> checking and unchecking individual scopes), and revoking tokens at the
> AS/RS.
>
> So the closest UX analog would probably be the wording displayed in an
> OAuth consent dialog, maybe something like:
>
>    - View [and prescribe] your medications
>    - View [and delete] your steps
>
>
>
>> If the *types* of Resource Sets and the allowed scopes are standardized
>> in advance (which UMA supports), then a mapping between Model 1 and
>> "vanilla" OAuth could be as simple as: "concatenate the UMA resource set
>> type followed by ':' followed by the UMA scope name" -- so for example, you
>> might derive an OAuth scope like "
>> https://openid.net/heart/resource-types/StepCounts:https://openid.net/heart/scopes/view".
>> Or under Model 2, the scopes could be reused directly (no mapping
>> required).
>>
>
> In what sense is "reuse" meant here? A coding model, or an architectural
> model, or a semantic model?... There are ways in which I could imagine a
> deep kind of semantic reuse being possible without concatenation tricks
> being necessary. However, not being a developer, I'm not sure if they match
> what you're thinking of.
>
> For example, in my previous response to the minutes email, I outlined how
> some APIs have implicit mappings between scopes and acceptable
> endpoints/resources to which they apply.
>
> Let's say (totally making this up) the FHIR has two endpoints, with one
> endpoint for medication records and one for fitness steps. There's an
> UMA-standardized resource type for each. There's "*https://www.hl7.org/fhir/rsrc/med.json
> <https://www.hl7.org/fhir/rsrc/med.json>*", with instances of it
> registered with scopes "*view*", "*download*", "*transmit*", and "*add*"
> (so some clients can insert new entries). Alice's medications might be in a
> resource something like "*/alice/meds*". (What's displayed in her AS
> dashboard might look a lot nicer than that.) And there's "*https://www.hl7.org/fhir/rsrc/step.json
> <https://www.hl7.org/fhir/rsrc/step.json>*", with instances of it
> registered with scopes "*view*", "*download*", "*transmit*", and "*chart*".
> Alice's steps might be in a resource like "/alice/steps".
>
> (If the scopes are in the form of URIs, they could be standardized to a
> further degree, in that a bunch of metadata could be pulled by the
> authorization server and used to present standard labels and icons, and
> other semantics could be linked to them.)
>
> If the very same API were OAuth-protected, with the very same resource
> endpoints, there might still be the same resource endpoints, with the same
> scopes, where three of them work on both resource types, "*add*" only
> works on "*med*", and "*chart*" only works on "*step*". These resources
> could still have a standardized meaning in terms of both resource naming
> and schema/format; there just would be nowhere to "hook" a standardized
> resource type URI into.
>
> Seen this way, the OAuth approach and the UMA approach are quite similar,
> differing only in the implicitness vs. explicitness of the resource set
> layer.
>
>
>>
>> Of course, some interesting things happen when we layer in details like...
>>
>> W*hat if Alice has access to multiple records (say, her own and her
>> mother's)?* In vanilla OAuth the binding of permissions to these records
>> is generally implicit. How should they play out in UMA? Under Model 1, we'd
>> probably see two more Resource Sets created ("Alice's Mom's Medications"
>> and "Alice's Mom's Steps"). Under Model 2, we'd probably see one more
>> Resource Set created ("Alice's Mom's FHIR Endpoint").
>>
>
> I've been doing some work around chained delegation of this sort. Indeed,
> these are separate records, and must remain that way. Alice may not have
> all the permissions over her mother's records that she has over her own!
> One way to present such "downstream" items is to present them under a
> separate "Shared With" area. And there are various ways to organize owned
> items, e.g. by who you tend to share them with or by function. In
> discussions with consumer IoT folks, it seems that smart light bulbs want
> to be gathered by "room".
>
> FWIW...
> _______________________________________________
> Openid-specs-heart mailing list
> Openid-specs-heart at lists.openid.net
> 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/20150616/86b20f2c/attachment.html>