OpenID, YADIS and Directed Identity

[Fen Laballme]

There are some interesting cryptographic properties that you can attach
to a set of nyms that can help to minimize data correlation (real person
triangulation).  I describe a few of these in an unpublished (and never
really completed) paper that may offer some useful insights at
<http://www.openprivacy.org/papers/200104-repcap.html>.  I reproduce the
most relevant section here:

Nym Properties

Privacy is ensured through the use of Nyms that have some very special
properties. Given

    * T = (Trusted) Parent Nym that can create Child Nyms
    * Ci, Cj and Ck are a set of child Nyms created by the same parent T
    * Ei is a child Nym from another parent

We can show:

   1. Ci, Cj, and Ck cannot prove that T is their Parent
   2. Ci, Cj, and Ck cannot prove they are siblings
   3. T can separately prove parenthood of Ci and/or Cj
         1. without also leaking information about any other siblings,
            e.g. Ck
   4. T can anonymously prove Ci and Cj are siblings
         1. without also leaking information about any other siblings,
            e.g. Ck
         2. and without leaking information that points to T as C's parent
   5. T cannot prove anything about E

=Fen


Drummond Reed wrote:
> Michael,
>
> We in XRI-land got the same feedback about the need for unidirectional
> identifiers for single sign-on last fall. And we came up with the same
> solution -- see my blog post at:
>
> http://www.equalsdrummond.name/index.php?p=56
>
> I think it applies to XRIs, URLs, or any other identifier technology. One
> caveat I'll point out, though: this technique only works if there are a
> sufficiently large enough set of identifiers issued by the identity
> service
> provider (in your example, idsrus.com). Otherwise you can be correlated at
> that level.
>
> =Drummond (http://xri.net/=drummond.reed)
>
> -----Original Message-----
> From: yadis-bounces at lists.danga.com [mailto:yadis-bounces at lists.danga.com]
> On Behalf Of Michael Graves
> Sent: Sunday, February 12, 2006 12:40 PM
> To: yadis at lists.danga.com
> Subject: Re: OpenID, YADIS and Directed Identity
>
> Johannes Ernst <jernst+lists.danga.com <at> netmesh.us> writes:
>
>>> ... in my
>>> scenario, you wouldn't enter "mart.whatever.com" at the initial
>>> login, screen.
>>> Instead you would only enter "whatever.com". At this point, then,
>>> the replying
>>> part only knows you are somehow attached to "whatever.com". You
>>> are then
>>> redirected (302) to whatever.com's login page. Unlike the current
>>> scenario,
>>> the identity server (whatever.com) has at this point no idea who
>>> you are, so
>>> instead of asking just for your password and presenting the "user"
>>> field
>>> already filled out, you would need to specify your user name at
>>> whatever.com's
>>> login screen as well.
>> Not necessarily. The identity server can have a cookie, shared only
>> with itself, that identifies who you are. So the sequence would be
>>
>> GET relying-party -> HTML form
>> POST relying party identity=whatever.com -> Redirect to whatever.com
>> GET whatever.com cookie=myid -> Redirect to whatever.com/myid
>> GET whatever.com/myid -> Redirect to relying party with signed URL
>> (if active session, otherwise ask for password first)
>>
>> P.S. No hunting party as long as everybody understands that this
>> is about something other than YADIS 1.0.
>>
>> Johannes Ernst
>> NetMesh Inc.
>>
>>
>>
>>
>> http://netmesh.info/jernst
>>
>>
>
> Hi Johannes!
>
> You're right, of course. My point was to emphasize the fact that the
> identity
> server has no idea who you are *from* the relying party, which means that
> subsequent provisioning of a generated ID won't provide a basis for
> correlation.
>
> If you are currently logged in to your identity server, your cookies can
> trigger "auto-fill" for your user name, when you are redirected
> to "whatever.com". That's a good point! So, ergonomically, you haven't
> given
> up anything important to the relying party yet, but when you arrive at
> whatever.com's site, it may know who you are through its own devices and
> thus
> you haven't really lost any momentum in the process.
>
> That's cool. I don't know what the real need/demand is for this
> functionality,
> but I do know that I've had to defend/address YADIS/OpenID not having it
> several times in the past few weeks. Again this is really an OpenID issue,
> not
> a YADIS one, and one we can take up later. This is an interesting
> "upgrade"
> to
> consider, though.
>
> -Mike
>
>
>
>
>
>