<html>
  <head>
    <meta content="text/html; charset=windows-1252"
      http-equiv="Content-Type">
  </head>
  <body bgcolor="#FFFFFF" text="#000000">
    Inline:<br>
    <br>
    <div class="moz-cite-prefix">On 10/19/2015 10:58 PM, Eve Maler
      wrote:<br>
    </div>
    <blockquote
cite="mid:CAMPbGmj2ZHnDhB24zv76vDXB5He_7WF9Q79N6cLb4vqyBWsGCg@mail.gmail.com"
      type="cite">
      <meta http-equiv="Content-Type" content="text/html;
        charset=windows-1252">
      <div dir="ltr">Hi Gunther-- Thanks for speaking up!
        <div><br>
        </div>
        <div>Interestingly, sharing with the hospital NPE could possibly
          feel just like sharing with a person with some identifier, if
          the resource owner is told what the hospital's identifier is
          and it's in a familiar format. Over time, we've all no doubt
          observed that email addresses, and things that look like them,
          are extremely powerful tools for representing identities --
          and even groups of them (because an email address can be an
          "alias"). As long as the address isn't literally a shared
          account, which is just bad news security-wise, maybe that
          pattern can be exploited somehow. ("Alice shares heart data
          stream with <a moz-do-not-send="true"
            href="mailto:cardio-unit@nyp.org">cardio-unit@nyp.org</a>"?)</div>
      </div>
    </blockquote>
    <br>
    It's important for us to remember that things that look like email
    addresses don't have to actually be email addresses. As long as it's
    an identifier that Alice can be told and can use, we're fine. I
    rather like the example above, since Dr. Bob can hand Alice
    <a class="moz-txt-link-rfc2396E" href="mailto:cardio-unit@nyp.org">"cardio-unit@nyp.org"</a> and she can probably generally understand
    that. On the technical side, we need to specify what that alias
    means, and this is likely to require profiling discovery. We took a
    step in the Privacy on FHIR project by saying it *was* an email
    address, using the following mechanism:<br>
    <br>
     - Alice enters <a class="moz-txt-link-rfc2396E" href="mailto:bob@nyp.org">"bob@nyp.org"</a> into her AS's policy page<br>
     - Alice's AS does a webfinger lookup on <a class="moz-txt-link-rfc2396E" href="mailto:bob@nyp.org">"bob@nyp.org"</a> to find the
    OpenID Connect Issuer for that identifier (let's say
    <a class="moz-txt-link-freetext" href="https://id.nyp.org/">https://id.nyp.org/</a> in this example)<br>
     - Alice's AS sets a policy that if someone comes in with the
    "email" field set to <a class="moz-txt-link-rfc2396E" href="mailto:bob@nyp.org">"bob@nyp.org"</a> from the <a class="moz-txt-link-rfc2396E" href="https://id.nyp.org">"https://id.nyp.org"</a>
    server we just discovered, and "email_verified" is set to "true",
    then to let them access things<br>
    <br>
    This works great for individuals where the identifier happens to
    match the email address (a common enough case for individuals of
    course), but falls over for any other attribute sets. It's important
    that whatever Alice enters be as simple as <a class="moz-txt-link-rfc2396E" href="mailto:bob@nyp.org">"bob@nyp.org"</a> or
    <a class="moz-txt-link-rfc2396E" href="mailto:cardio-unit@nyp.org">"cardio-unit@nyp.org"</a> to kick off the whole process, otherwise it's
    far too much work for the end users. Webfinger is the right starting
    point, but I believe we need something beyond that.<br>
    <br>
    <blockquote
cite="mid:CAMPbGmj2ZHnDhB24zv76vDXB5He_7WF9Q79N6cLb4vqyBWsGCg@mail.gmail.com"
      type="cite">
      <div dir="ltr">
        <div><br>
        </div>
        <div><b>"T"-for-technical observation coming...</b> Another
          perhaps interesting observation (knowing this will probably
          annoy Justin until some perfected version of UMA shows up on
          the horizon :-): Agents of NPE RqPs, such as all the employees
          of one hospital, will probably all share the same mechanism
          for identifying themselves to Alice's AS, at least for the
          purpose of the initial consent to share claims with it. (This
          is the "AAT issuance" flow.) If it were possible to make one
          additional (huge?) simplifying assumption, which is that Dr.
          Bob <i>will use</i> <i><b>Alice's</b></i> <i>AS</i> whenever
          he wants to further share her resource with some
          proxy-for-Bob, and proxy Charlie only <i>uses</i> <i>Alice's</i>
          <i>AS</i> whenever he wants to share Alice's resource with yet
          another hospital employee down the line, then we can easily do
          this with "today's UMA", and Alice will easily be able to
          inspect the sharing chain from a central location and cut off
          access if she sees that the ability is being abused.</div>
        <div><br>
        </div>
        <div>If Alice were to share with <a moz-do-not-send="true"
            href="mailto:cardio-unit@nyp.org">cardio-unit@nyp.org</a>,
          then -- I think -- whatever/whoever controls that alias would
          indeed have to use Alice's AS to set/generate policies for
          delegation chaining, in order for it to work for the next hop.</div>
      </div>
      <div class="gmail_extra"><br clear="all">
      </div>
    </blockquote>
    <br>
    I don't believe this is a simplifying assumption that we can
    reasonably make, thanks to the caching issue brought up on today's
    call. Dr. Bob will be at the junction between Alice's resource and
    his own data set. Authorization decisions that he makes will not be
    for Alice's resource, but they will be for *his copy* of *her
    resource*. This is a key distinction that we can't ignore. Once the
    data is copied, Alice effectively has no control over what happens
    to it from a technical level. This is the nature of digital
    information, and it's up to the business and legal layers to do
    something about it. It's possible to reach back into the technical
    space with a component like a Consent Receipt to represent rights
    and responsibilities at these other layers, but Alice still has to
    trust Dr. Bob to actually do what he said he'd do here, and I don't
    think there's any way around that. Previous attempts to control
    digital data once it's been released have been known as "DRM" and
    are technologically untenable. <br>
    <br>
    Forcing Bob to go back to Alice's AS to set further policies has
    other problems as well, as now Bob needs a login on the AS that's
    capable of setting policies. In current UMA Bob also needs an
    account capable of issuing OAuth tokens (the AAT) but that's a
    separately broken thing in UMA that we should not rely on. Not the
    least of which reason being that I want to tear that part of the
    spec out as soon as reasonably possible, so we shouldn't start
    building on it. There are other issues as well. Regardless, we don't
    need that assumption to accomplish what we want to do, and it's
    actively harmful and limiting if we make it.<br>
    <br>
     -- Justin<br>
    <br>
    <blockquote
cite="mid:CAMPbGmj2ZHnDhB24zv76vDXB5He_7WF9Q79N6cLb4vqyBWsGCg@mail.gmail.com"
      type="cite">
      <div class="gmail_extra">
        <div>
          <div class="gmail_signature">
            <div dir="ltr">
              <div>
                <div dir="ltr">
                  <p><b>Eve Maler<br>
                    </b>ForgeRock Office of the CTO | VP Innovation
                    & Emerging Technology<br>
                    Cell +1 425.345.6756 | Skype: xmlgrrl | Twitter:
                    @xmlgrrl<br>
                    Join our <a moz-do-not-send="true"
                      href="http://forgerock.org/openuma/"
                      target="_blank">ForgeRock.org OpenUMA</a>
                    community!</p>
                </div>
              </div>
            </div>
          </div>
        </div>
        <br>
        <div class="gmail_quote">On Mon, Oct 19, 2015 at 3:13 PM, Meyer,
          Gunther <span dir="ltr"><<a moz-do-not-send="true"
              href="mailto:Gunther.Meyer@allscripts.com" target="_blank">Gunther.Meyer@allscripts.com</a>></span>
          wrote:<br>
          <blockquote class="gmail_quote" style="margin:0 0 0
            .8ex;border-left:1px #ccc solid;padding-left:1ex">
            <div link="blue" vlink="purple" lang="EN-US">
              <div>
                <p class="MsoNormal"><span
style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1f497d">Hi
                    All!</span></p>
                <p class="MsoNormal"><span
style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1f497d"> </span></p>
                <p class="MsoNormal"><span
style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1f497d">First
                    off, thank you for letting me participate in this
                    discussion.  And please excuse the Newbie questions
                    or suggestions!</span></p>
                <p class="MsoNormal"><span
style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1f497d"> </span></p>
                <p class="MsoNormal"><span
style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1f497d">From
                    a sharing of information point of view, if I put
                    myself in the shoes of a patient, I see that sharing
                    information with an individual makes sense, with the
                    following provisions</span></p>
                <p><span
style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1f497d"><span>1.<span
                        style="font:7.0pt "Times New Roman"">      
                      </span></span></span><span
style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1f497d">If
                    the provider is not available, someone else that is
                    their proxy must be able to access it.</span></p>
                <p><span
style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1f497d"><span>2.<span
                        style="font:7.0pt "Times New Roman"">      
                      </span></span></span><span
style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1f497d">I
                    may not know the provider, for example if this is
                    the first time at the practice (might be an edge
                    case?)</span></p>
                <p class="MsoNormal"><span
style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1f497d"> </span></p>
                <p class="MsoNormal"><span
style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1f497d">As
                    to the “group” concept, I did not want to get into
                    the middle of a debate.  I just wanted to point out
                    that larger practices or hospital often consider
                    sharing in terms of groups of users rather than
                    individuals, and would therefore be more comfortable
                    with a sharing model that involved groups. Therefore
                    it is just something we need to consider (though we
                    might reject is as impractical)</span></p>
                <p class="MsoNormal"><span
style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1f497d"> </span></p>
                <p class="MsoNormal"><span
style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1f497d">Could
                    someone please share the link to the BLT
                    presentation, that sounded very interesting.</span></p>
                <p class="MsoNormal"><span
style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1f497d"> </span></p>
                <p class="MsoNormal"><span
style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1f497d">Also,
                    I think that the point that the application, C,
                    might have to be better defined seems to have merit,
                    unless the patient just authorizes the Hospital?</span></p>
                <p class="MsoNormal"><span
style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1f497d"> </span></p>
                <p class="MsoNormal"><span
style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1f497d">Finally,
                    I think that the patient sharing with an individual
                    doctor is interesting if the user’s intent can be
                    preserved during the import process.  I think very
                    few users, however, will express much more than the
                    default “share with anyone as needed for treatment”
                    consent.  Therefore, once the information is shared
                    with NYP, as discussed, it will be imported and then
                    the organizational security and policies will kick
                    in.</span></p>
                <p class="MsoNormal"><span
style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1f497d"> </span></p>
                <p class="MsoNormal"><span
style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1f497d">Regards</span></p>
                <p class="MsoNormal"><span
style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1f497d"> </span></p>
                <p class="MsoNormal"><span
style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1f497d">Gunther</span></p>
                <p class="MsoNormal"><span
style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1f497d"> </span></p>
                <p class="MsoNormal" style="line-height:12.0pt"><b><span
style="font-size:9.0pt;font-family:"Arial",sans-serif;color:#5b8f22">Gunther
                      Meyer</span></b><span
style="font-size:9.0pt;font-family:"Arial",sans-serif;color:#707176">
                    |
                  </span><b><span
style="font-size:9.0pt;font-family:"Arial",sans-serif;color:#616365">Architect<br>
                    </span></b><b><span
style="font-size:9.0pt;font-family:"Arial",sans-serif;color:#5b8f22">Allscripts</span></b><span
style="font-size:9.0pt;font-family:"Arial",sans-serif;color:#707176">
                    | 8529 Six Forks Road | Raleigh, NC |27615</span></p>
                <p class="MsoNormal"><span
style="font-size:9.0pt;font-family:"Arial",sans-serif;color:#616365"> <br>
                    <a moz-do-not-send="true" href="tel:919.329.1466"
                      value="+19193291466" target="_blank">919.329.1466</a></span><span
style="font-size:9.0pt;font-family:"Arial",sans-serif;color:black">|
                  </span><span
style="font-size:9.0pt;font-family:"Arial",sans-serif;color:green">P</span></p>
                <p class="MsoNormal"><span
style="font-size:9.0pt;font-family:"Arial",sans-serif;color:#616365"><a
                      moz-do-not-send="true" href="tel:919.457.4466"
                      value="+19194574466" target="_blank">919.457.4466</a></span><span
style="font-size:9.0pt;font-family:"Arial",sans-serif;color:black">|
                  </span><span
style="font-size:9.0pt;font-family:"Arial",sans-serif;color:green">F</span></p>
                <p class="MsoNormal"><span
style="font-size:9.0pt;font-family:"Arial",sans-serif;color:#616365"><br>
                  </span><u><span
style="font-size:9.0pt;font-family:"Arial",sans-serif;color:#5b8f22"><a
                        moz-do-not-send="true"
                        href="mailto:gunther.meyer@allscripts.com"
                        target="_blank"><a class="moz-txt-link-abbreviated" href="mailto:gunther.meyer@allscripts.com">gunther.meyer@allscripts.com</a></a></span></u><span
style="font-size:9.0pt;font-family:"Arial",sans-serif;color:black">
                    |
                  </span><span
style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1f497d"><a
                      moz-do-not-send="true"
                      href="http://www.allscripts.com/" target="_blank"><span
style="font-size:9.0pt;font-family:"Arial",sans-serif;color:#5b8f22"><a class="moz-txt-link-abbreviated" href="http://www.allscripts.com">www.allscripts.com</a></span></a></span><span
style="font-size:9.0pt;font-family:"Arial",sans-serif;color:green">
                  </span></p>
                <p class="MsoNormal"><span
style="font-size:9.0pt;font-family:"Arial",sans-serif;color:#4d4f53">Corporate
                    Headquarters l 222 Merchandise Mart Plaza l 20<sup>th</sup>
                    Floor l Chicago, IL l 60654</span><span
style="font-size:9.0pt;font-family:"Arial",sans-serif;color:green"></span></p>
                <p class="MsoNormal"><b><span
style="font-size:9.0pt;font-family:"Arial",sans-serif;color:#76923c"> </span></b></p>
                <p class="MsoNormal"
                  style="margin-bottom:5.25pt;line-height:15.0pt"><u><span
style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#76923c"><span
                        style="text-decoration:none"> </span></span></u></p>
                <p class="MsoNormal"><b><span
style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1f497d"> </span></b></p>
                <p class="MsoNormal"><span
style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1f497d"> </span></p>
                <p class="MsoNormal"><span
style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1f497d"> </span></p>
                <p class="MsoNormal"><b><span
                      style="font-size:11.0pt;font-family:"Calibri",sans-serif">From:</span></b><span
style="font-size:11.0pt;font-family:"Calibri",sans-serif">
                    Openid-specs-heart [mailto:<a moz-do-not-send="true"
href="mailto:openid-specs-heart-bounces@lists.openid.net"
                      target="_blank">openid-specs-heart-bounces@lists.openid.net</a>]
                    <b>On Behalf Of </b>Eve Maler<br>
                    <b>Sent:</b> Monday, October 19, 2015 4:00 PM<br>
                    <b>To:</b> <a moz-do-not-send="true"
                      href="mailto:openid-specs-heart@lists.openid.net"
                      target="_blank">openid-specs-heart@lists.openid.net</a><br>
                    <b>Subject:</b> [Openid-specs-heart]
                    "Individual-to-NPE" sharing episodes in UMA use
                    cases, and "design pattern" solution options</span></p>
                <div>
                  <div class="h5">
                    <p class="MsoNormal"> </p>
                    <div>
                      <p class="MsoNormal">I promised to write this up,
                        and hopefully I'll make it before the deadline
                        of today's call.</p>
                      <div>
                        <p class="MsoNormal"> </p>
                      </div>
                      <div>
                        <p class="MsoNormal">The subject line introduces
                          what I hope will be useful consistent wording
                          for discussing these sorts of topics. Some of
                          our UMA use cases include episodes of
                          party-to-party resource sharing that involve a
                          resource owner who is an individual (say, a
                          patient or consumer), and a requesting party
                          that <b>is, or is the agent of,</b> a
                          "non-person entity" or NPE, such as a
                          hospital, government agency, or company.</p>
                      </div>
                      <div>
                        <p class="MsoNormal"> </p>
                      </div>
                      <div>
                        <p class="MsoNormal">Staying entirely within the
                          confines of the UMA protocol, a number of
                          different "design patterns" could be chosen
                          for deployment. Agreeing on which reasons to
                          use which patterns, and locking down any areas
                          of variability, could help make systems
                          interoperate with each other. The UMA
                          protocol, in fact, expects such variability
                          and recommends profiling to improve
                          interoperability. Thus, it seems a good idea
                          for us to figure out how much such types of
                          interop are in scope for us, and likely <b>do
                            some profiling in these areas</b>.</p>
                      </div>
                      <div>
                        <p class="MsoNormal"> </p>
                      </div>
                      <div>
                        <p class="MsoNormal">Here are four patterns I
                          can think of:</p>
                      </div>
                      <div>
                        <ol start="1" type="1">
                          <li class="MsoNormal">
                            <b>Individual-to-agent-of-NPE</b>: Alice the
                            individual RO shares with "the individual
                            RqP who can prove they control the
                            identifier 'Dr. Bob'" (possible also
                            constraining the client in use as well --
                            we'll leave that part out for this
                            analysis).</li>
                          <li class="MsoNormal">
                            <b>Individual-to-NPE</b>: Alice the
                            individual RO shares with "the NPE RqP that
                            can prove they control the identifier 'New
                            York Presbyterian Hospital'". Some process
                            yet to be determined, possibly involving
                            "chained delegation", ensures that Dr. Bob
                            and possibly others who work for NYP get
                            access thereafter.</li>
                          <li class="MsoNormal">
                            <b>Individual-to-role</b>: Alice the
                            individual RO shares with "any RqP who can
                            prove they have been assigned the role
                            'works for NYP'".</li>
                          <li class="MsoNormal">
                            <b>Individual-to-individual</b>: Alice the
                            individual RO shares with "the individual
                            RqP who can prove they control the
                            identifier 'bob@gmail'" (whom she knows is
                            her doctor because he provisioned her with
                            that gmail handle). Bob might do "chained
                            delegation" to share the resource with
                            himself as an employee of NYP.</li>
                        </ol>
                      </div>
                      <div>
                        <p class="MsoNormal">The reason interop
                          questions arise is because the process of UMA
                          trust elevation involves things like
                          claims-gathering and possibly step-up
                          authentication, and the policy-setting options
                          presented to Alice (which are out of band of
                          UMA, but nonetheless...) need to be driven by
                          these requirements. The ability of the
                          requesting sides to respond appropriately will
                          be triggered off of expectations about what
                          they'll be asked to cough up for trust
                          elevation.</p>
                      </div>
                      <div>
                        <p class="MsoNormal"> </p>
                      </div>
                      <div>
                        <p class="MsoNormal">Each pattern has pros and
                          cons. Briefly:</p>
                      </div>
                      <div>
                        <ul type="disc">
                          <li class="MsoNormal">
                            The one I'm least enamored of is #3;
                            enterprise access control has had so much
                            trouble with RBAC, so can we expect adding
                            UMA to help? :-)</li>
                          <li class="MsoNormal">
                            Chained delegation can be very powerful. In
                            environments where everybody uses the same
                            UMA authorization server, a number of nice
                            value-add features can be supported, but
                            they tend to break down (at least with UMA
                            V1.0.x) when you add the ability for every
                            RO to choose their own AS.</li>
                          <li class="MsoNormal">
                            I worry about sharing with individual
                            doctors. It's very expedient, so people will
                            tend to do it as a path of least resistance
                            (think Google Apps!). And sometimes maybe
                            it's the right answer, particularly if
                            "chained delegation" can allow Alice to
                            track where her resource has been shared
                            further. But what if Dr. Bob leaves the
                            hospital/practice/whatever? Is this always
                            the right answer?</li>
                          <li class="MsoNormal">
                            Sharing with an NPE sounds elegant -- it's
                            what a recent POC of my acquaintance did.
                            But the "process yet to be determined"
                            mentioned above wasn't actually determined
                            yet, so there's that. :-) And you have the
                            problem of a system administrator who has
                            privileged identity credentials to the NPE
                            account -- as always -- having the key to a
                            pretty valuable kingdom. Maybe a cool
                            mitigation of this risk is to go with
                            sharing with individuals and tracking
                            sharing chains?</li>
                        </ul>
                      </div>
                      <div>
                        <div>
                          <div>
                            <div>
                              <div>
                                <div>
                                  <p><b>Eve Maler<br>
                                    </b>ForgeRock Office of the CTO | VP
                                    Innovation & Emerging Technology<br>
                                    Cell <a moz-do-not-send="true"
                                      href="tel:%2B1%20425.345.6756"
                                      value="+14253456756"
                                      target="_blank">+1 425.345.6756</a>
                                    | Skype: xmlgrrl | Twitter: @xmlgrrl<br>
                                    Join our <a moz-do-not-send="true"
                                      href="http://forgerock.org/openuma/"
                                      target="_blank">ForgeRock.org
                                      OpenUMA</a> community!</p>
                                </div>
                              </div>
                            </div>
                          </div>
                        </div>
                      </div>
                    </div>
                  </div>
                </div>
              </div>
            </div>
          </blockquote>
        </div>
        <br>
      </div>
      <br>
      <fieldset class="mimeAttachmentHeader"></fieldset>
      <br>
      <pre wrap="">_______________________________________________
Openid-specs-heart mailing list
<a class="moz-txt-link-abbreviated" href="mailto:Openid-specs-heart@lists.openid.net">Openid-specs-heart@lists.openid.net</a>
<a class="moz-txt-link-freetext" href="http://lists.openid.net/mailman/listinfo/openid-specs-heart">http://lists.openid.net/mailman/listinfo/openid-specs-heart</a>
</pre>
    </blockquote>
    <br>
  </body>
</html>