[Openid-specs-heart] An approach to data portability for the RO's policies

Eve Maler eve.maler at forgerock.com
Wed Aug 31 13:13:14 UTC 2016


An UMA RS does not understand, nor have to understand, policy; no entity
needs to understand policy in order to get UMA benefits. An RS only has to
understand how to interact with the UMA protection API and with UMA
clients. That is still true, even if other ecosystem players (AS's today,
maybe other peripheral players tomorrow) add value through policy
manipulation, at an API I've just sketched that isn't even standardized
today and could be complex and multi-natured. I have no idea how you
imagine such an API destroying UMA's salutary properties given that *it's
not even part of UMA and we can't prevent anyone from coming up with it and
UMA-protecting it anyway*.

You know what, never mind. Sorry I tried to explain how policy could be
portable.


*Eve Maler*ForgeRock Office of the CTO | VP Innovation & Emerging Technology
Cell +1 425.345.6756 | Skype: xmlgrrl | Twitter: @xmlgrrl
*ForgeRock Summits and UnSummits* are coming to
<http://summits.forgerock.com/> *London and Paris!*

On Wed, Aug 31, 2016 at 5:39 AM, Adrian Gropper <agropper at healthurl.com>
wrote:

> Debbie,
>
> My stance is simply patient-centered. There's nothing patient-centered
> about not giving the patient a choice of _one thing_, anything, that they
> can specify when an RS claims HEART compliance.
>
>    - Can the patient specify a notification email address when a new
>    Client is registered?
>    - Can she specify a notification when a new RqP seeks access?
>    - Can she register another RS to use the same AS that she is being
>    forced to use?
>    - Does HEART offer a solution to Alice's multiple portals problem?
>    - What does HEART mean if Alice can't specify the AS?
>
> I think you may be confusing the UI issue with how UMA works. I have no
> problem with providing a UI for policies at the AS as long as we're clear
> that Alice must be able to specify the AS. That would solve the problem you
> seem to be trying to solve because the RS would be able to "bolt on" a UI
> to capture Alice's policy and then send that policy directly to Alice's AS
> using OAuth. In that instance, the thing you're calling a "privacy manager"
> is not an AS in the HEART sense although it can have UMA functionality in
> cases where Alice does not have an AS of her own choice.
>
> Let's keep going with this "privacy manager" concept and take it to the
> next level where the "privacy manager" runs a full UMA AS and Alice tries
> to register another resource server. Let's assume the "privacy manager AS"
> is run by the Veterans Health Administration and the new resource server
> that Alice wants to register holds her social media data and is not HIPAA
> covered and maybe not even in US jurisdiction. Now, the VA as AS operator
> is forced to be responsible for issuing access tokens to some RS in some
> other country that holds personal data about Alice that's not even close to
> their responsibilities. What's the likelihood of the VA doing that? Is
> there any amount of XACML that can protect the VA under these
> circumstances? Will the VA accept logins and credentials from would-be Bob
> RqPs that come asking for access?
>
> Under this "policies on the wire" construct, the VA would have a choice to
> say to Alice: "You can't register this particular RS here." please use
> another AS. At that point, Alice has two ASs and the exchange of policies
> between them via an interface does not help Alice or make HEART work. How
> does the VA AS work with the other AS?
>
> So, to put this in terms that "privacy manager' vendors (and RS customers)
> can understand, they can go into the UMA Authorization Business and we hope
> they do. Their AS policy interface can be tightly coupled to a particular
> RS in order to provide some usability benefits because the policy UI is
> presented to the patient in the context of the RS. That is a real value
> add. Michael Chen and I implemented this feature in the current HIE of One
> demo between Alice's HIE of One AS and Alice's pNOSH PHR because it
> presents Alice's UI in the PHR context so she can see what information
> would be released as she makes her elections. However, this does not avoid
> the RS offering Alice the choice of an AS that she specifies. The
> HEART-compliant "privacy manager" will need to honor the AS whether it's
> specified by Alice or built and run by the privacy manager vendor and their
> customer.
>
> Adrian
>
>
>
>
> On Wed, Aug 31, 2016 at 7:38 AM, Debbie Bucci <debbucci at gmail.com> wrote:
>
>> I don't see it that way.  Your stance seems to be all or nothing.   How
>> can you encourage an ecosytem to grow by being so rigid?
>>
>> Alice would not be forced to share anything.   I see it more of enabling
>> the bolt on of privacy managers - potentially minimizing the UI placed on
>> RS   and easing the burden of data entry for the patient.  There are policy
>> requirement outside the UMA protocol that will need to be satisfied (thank
>> you Luis for that clarification/thought).
>>
>> Alice can build/run AND outsource.
>>
>
>
>
> --
>
> 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/
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openid.net/pipermail/openid-specs-heart/attachments/20160831/f69fa3e5/attachment.html>


More information about the Openid-specs-heart mailing list