I have very serious concerns about your perspective.  In my world, what a relying party needs to know is seldom "absolute identity" but simply what "is true about the entity" in order to make an access management decision.  The latter can be achieved in many ways.

"Absolute identity" is whatever it takes to distinguish one entity from all others.  DNA works pretty well.  Driver's license, not so well.

If "absolute" identity (AId) is able to be determined with relative ease, then we will have lost every vestige of privacy.  Period.  It will be used by every RP from the NSA to Google and Facebook.  Recall how misused SSN became.  Everything will be tied to AId.  Everything you do will be correlatible.

AId may be important in arresting someone but only if guilt was established in a way that tied to that identity.  Almost everything else between a Subject and RP is relative, or contextually dependent.  If I withdraw money fro an ATM, the "relative identity" is that I own that account number.  If I go to a hospital for treatment, the relative identity is that I am the person with patient # 11235813.

Perhaps we need to drop the use of "identity" altogether and use terms that are less ambiguous.  The notion of an "Identity Service Provider" could become "Subject Information Provider" (SIP).  AId would become "Globally Unique Identifier" (GUI).

My ideal model of the person centric ecosystem is one in which I have strong, reliable credentials (plural) with which I can associate authoritative information acquired and held by one or more SIP(s).  The set of information known to any given SIP would not necessarily be known by other SIPs.  SIPs would be forbidden by law from colluding.  I am given complete control of the release of my SIP information to RPs with which I wish to interact.  None of this requires a GUI.

David


On Mar 26, 2012, at 1:47 PM, Joanne Knight wrote:

There are two point I would pick up here.
 
1) If we define Identity Attribute as the values needed to establish that an individual entity in unique then adding attributes such as Student would mean an unnecessary number of attributes would need to be collected to achieive uniqueness.
 
2) As per previous email, we have moved further to a point where we consider Role or Context separately from unique identity. For example in an Education system one unique identity may have more than one role e.g. Student and Teacher. The identity is the same for both but the role they perform in different situations changes. We would attach rights and permissions to the combination of identity and role (personna) not to the identity on its own. Identifiers, facts and preferences may also be different depending on the role. e.g. Joe Bloggs teaches Mathematics but is learning how to paint. In his personna of Mathematics teacher his qualifications and use of the title Professor are relevant, when he is learning to paint he no longer wishes to be referred to as Professor and his mathematic and teaching qualifications are also unlikely to be relevant to his learning or ability to paint.
 
Don't we want to move to an environment where a person can establish their unique identity and attach it to a credential that allows them to reuse this identity when registering with any organisation. If we encumber that identity with role information that is only relevant to one organisation or system context then I would suggest we would be designing a global service centric data store rather than a person centric ecosystem.
 
Just my thoughts
 
From: dg-am-bounces@kantarainitiative.org [mailto:dg-am-bounces@kantarainitiative.org] On Behalf Of David L. Wasley
Sent: Friday, 23 March 2012 8:02 a.m.
To: dg-am@kantarainitiative.org
Subject: Re: [DG-AM] definition of Identity Attribute for the report
 
Some background and a different perspective.
 
The original IAF was developed from the original NIST 800-63.  The notion of "identity" in 800-63 was basically "a name and something added to make it unique" which is where the definition Ken found came from.  The world has moved quite far from that (naive) notion.
 
In the identity federation with which I am most familiar (InCommon), we consider "identity attributes" to be potentially "anything that is true about a given entity."  Identifiers, facts, preferences, etc.  For example, "student" is an important attribute that supports access to services and resources restricted to "students."  The basic set of attributes we use is defined in the eduPerson Object Class (http://middleware.internet2.edu/eduperson/).  
 
Another set of attributes that we have discussed but not implemented would provide information that can help a RP/SP display information to a user, for example "visually impaired" -->> "increase font size," or "color blind" or "deaf", etc.
 
In this broader notion of "attribute" there are many different kinds:
 - some things are unique to the particular entity; others are shared with other entities
 - some are assigned by an SOA, e.g., passport number; others are self selected, e.g., nickname or display name.
 - some are transient, e.g., "student"; others are permanent and/or never reassigned, e.g., some identifiers.
 - the degree of authoritativeness of any attribute is determined by how it is acquired by the ISP and how current it is, etc.
 - not all attributes will be available from any one ISP, certainly not with the same degree of assurance
 - etc.
 
We encourage RP/SPs to request the minimum set of attributes that they need in order make an access decision.
 
            David
 
On Mar 22, 2012, at 10:47 AM, Salvatore D'Agostino wrote:


Thanks Ken,
 
OK so if we move away from ITU f, toward  "Identity Attribute is information that contributes to establishing the identity (a unique name) of a single person?"
 
Yes attributes can support a “higher level of authN” but also are related to authZ independent of or in combination with name or identifier.   It depends on the attribute types, so might we expand this to include “.. contributes to establishing the identity (unique name) and “permissions/privileges/claims” of an individual”
 
Not sure what the actual word is there and put these 3 in as example/suggestion.
 
From: dg-am-bounces@kantarainitiative.org [mailto:dg-am-bounces@kantarainitiative.org] On Behalf Of Dagg, Kenneth
Sent: Thursday, March 22, 2012 12:56 PM
To: dg-am@kantarainitiative.org
Subject: [DG-AM] definition of Identity Attribute for the report
 
I checked for the term Identity Attribute in the IAF Glossary and did not find it.  As such, I did not send a note to the IAWG.
 
However, the following terms are in the glossary:
 
* Attribute - a property associated with an individual
* Identity - a unique name for a single person. Because a person’s legal name is not necessarily unique, identity must include enough additional information (for example, an address or some unique identifier such as an employee or account number) to make a unique name.
* Identification - Process of using claimed or observed attributes of an individual to infer who the individual is.
* Identity Proofing - The process by which identity related information is validated so as to identify a person with a degree of uniqueness and certitude sufficient for the purposes for which that identity is to be used.
 
The AMDG report currently defines Identity Attribute as Information bound to a subject identity that specifies a characteristic of the subject.
 
I suggest that this definition is not in alignment with the definitions contained in the IAF glossary. While I have nothing against the definitions contained in ITU-T X.1252 I would suggest that we remain consistent and aligned with KI definitions. I believe the following would be more aligned, "Identity Attribute is information that contributes to establishing the identity (a unique name) of a single person?"
 
Comments? Or reasons not to use this definition (other than it’s not the ITU definition)?
 
BTW: I have updated the report. I added a glossary and some text about RP requirements.  I also took the opportunity to align the recommendations at the start of the report with the recommendations at the end.
 
Ken
 
 
Kenneth Dagg
Senior Project Co-ordinator | Coordonnateur de projet supérieur
Security and Identity Management | Sécurité et gestion des identités
Chief Information Officer Branch | Direction du dirigeant principal de l'information
Treasury Board of Canada Secretariat | Secrétariat du Conseil du Trésor du Canada
Ottawa, Canada K1A 0R5
Kenneth.Dagg@tbs-sct.gc.ca
Telephone | Téléphone 613-957-7041 / Facsimile | Télécopieur 613-954-6642 / Teletypewriter | Téléimprimeur 613-957-9090
Government of Canada | Gouvernement du Canada

 
 
 
_______________________________________________
DG-AM mailing list
DG-AM@kantarainitiative.org
http://kantarainitiative.org/mailman/listinfo/dg-am
 
====
CAUTION:  This email message and any attachments contain information that may be confidential and may be LEGALLY PRIVILEGED. If you are not the intended recipient, any use, disclosure or copying of this message or attachments is strictly prohibited. If you have received this email message in error please notify us immediately and erase all copies of the message and attachments. Thank you.
====