OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.

 


Help: OASIS Mailing Lists Help | MarkMail Help

xacml message

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]


Subject: Re: [xacml] Minutes of 12 Feb 2004 Focus Group


No, it took 1 and 1/2 hours :-)

Anne

On 12 February, Polar Humenn writes: Re: [xacml] Minutes of 12 Feb 2004 Focus Group
 > From: Polar Humenn <polar@syr.edu>
 > To: Anne Anderson <Anne.Anderson@Sun.COM>
 > Cc: XACML TC <xacml@lists.oasis-open.org>
 > Subject: Re: [xacml] Minutes of 12 Feb 2004 Focus Group
 > Date: Thu, 12 Feb 2004 15:28:15 -0500 (EST)
 > 
 > 
 > Wow! Did you three discuss all this stuff in a hour?
 > -Polar
 > 
 > On Thu, 12 Feb 2004, Anne Anderson wrote:
 > 
 > > Minutes from the 12 February 2004 XACML Focus Group
 > >
 > > Present:
 > >   Anne Anderson,
 > >   Tim Moses,
 > >   Simon Godik
 > >
 > > Agenda: Continuing discussing XACML 2.0 work items
 > >
 > > Note 1: Anne will miss the Feb. 19 TC meeting.  On vacation.
 > > Note 2: Tim will put out a new draft once ConditionReference is
 > >   settled.
 > >
 > > XACML 2.0 Work Items discussed
 > >
 > > 7. ConditionReference
 > >
 > >   Wait for this item to receive more discussion on the list
 > >   before trying to resolve this.
 > >
 > > 9. Policies referring to hierarchical resources
 > >
 > >   RECOMMENDATION: Include items from #42 that actually refer to
 > >   the policy side.
 > >
 > > 29. Policy Authority Delegation
 > >
 > >   Should be on the open item list and should be a major topic to
 > >   be resolved at the next F2F.
 > >
 > > 31. Attribute Issuer as Subject
 > >
 > >   This really depends on the administrative policy model, which
 > >   is not settled.  This might get resolved at the next F2F.
 > >
 > >   ACTION: request use cases from those interested in this item.
 > >   Plan to agree on a model at the next F2F.  This will require
 > >   people to do homework ahead of time.
 > >
 > > 32. Standardize naming to specify rules for requestor's authz policy
 > >
 > >   Why does the naming have to switch?  Can't we just agree that
 > >   the naming used on out-bound policies must be consistent with
 > >   the naming used on the in-bound policies to which the out-bound
 > >   policy would apply.  This came up in WSPL, and was handled via
 > >   profile description and specification.
 > >
 > >   Might be like in RPC, where the two sides may have different
 > >   names for parameters, but they are consistent so long as in
 > >   same order and same data type.
 > >
 > >   ACTION: need Frank to explain his use case in a teleconference,
 > >   not just in an e-mail.
 > >
 > > 36. Check for requester authorized to ask for authz decision
 > >
 > >   One use case here is where a PDP sees a policy reference in its
 > >   policies, knows there is another PDP that can handle that
 > >   policy, passes along the Request Context to that PDP and asks
 > >   for a decision.  Another use case is equivalent to passing a
 > >   policy along with the request to a PDP, and asking for an
 > >   evaluation based on that policy.
 > >
 > >   We do not currently have a mechanism for passing a policy along
 > >   with a Request Context.
 > >
 > >   RECOMMENDATION: This should be an aspect of our Administrative
 > >   Policy, and should be resolved at the next F2F as part of the
 > >   overall Administrative Policy framework.
 > >
 > > 37. Multiple <AttributeValue> elements for single <Attribute> in Request
 > >
 > >   Have we addressed all of Rebekah's questions?
 > >
 > >   We think this item is now resolved/closed as in Draft 5,
 > >   subject to re-opening if Rebekah has another issue.
 > >
 > > 38. Policies for the Administration of XACML Policies
 > >
 > >   RECOMMENDATION: Close this as duplicate of #29: Policy
 > >   Authority Delegation.
 > >
 > > 42. Requests asking for access to multiple elements in a hierarchical resource
 > >
 > >   Note that we already have mechanism for asking for multiple
 > >   things in a hierarchy from the requesters point of view.  That
 > >   is closed.
 > >
 > >   Still issue on how to handle cases like "read" in a file system
 > >   requiring "execute" permission on all elements higher in the
 > >   hierarchy.  This belongs under #9.
 > >
 > >   RECOMMENDATION: Close this one.
 > >
 > > 45. Fix AttributeAssignment example in Section 4.2.4.3 (Rule 3)
 > >
 > >   RECOMMENDATION: Close, since no objections to current draft.
 > >
 > > 46. Status detail for missing attributes
 > >
 > >   Still awaiting "missing attribute" schema from Seth.  Could
 > >   solve by using AttributeDesignator?  Can't use <Attribute>,
 > >   since it requires a value.  May just want to specify
 > >   AttributeId, Issuer, DataType.
 > >
 > >   RECOMMENDATION: Seth, get busy.
 > >
 > > 51. Clarify obligations: "fulfill" vs. "understand"
 > >
 > >   Tim has already clarified this in the current draft.
 > >
 > >   RECOMMENDATION: close as fixed in current draft.
 > >
 > > 52. New section explaining not backwards compatible and listing changes
 > >
 > >   Full text requires resolution of other issues.  Bill on the
 > >   hook eventually.
 > >
 > > 55. Converge SAML and XACML Attribute schema definitions
 > >
 > >   Need algorithm mapping between SAML Attribute and XACML
 > >   Attribute.  Some issues are keeping Issuer syntax compatible
 > >   and handling two-part SAML Attribute names (domain plus id) and
 > >   one-part XACML Attribute.  Might be able to handle some of this
 > >   using AttributeSelector, and inserting SAML Attribute Assertion
 > >   as AttributeValue with DataType of saml:Assertion or
 > >   saml:Attribute.
 > >
 > > 58. Standard hierarchy schemas
 > >
 > >   We need say to associate attributes with each node in the
 > >   hierarchy.  Look at W3C ResourceDescriptionFramework: we don't
 > >   know if this applies, but the name sounds promising.
 > >
 > > 59. Define standard "role" subject attribute
 > >
 > >    Add following to Appendix B.5 "Subject attributes":
 > >
 > >     This identifier indicates a role associated with the subject.
 > >     Values of this attribute SHOULD be of type
 > >     "http://www.w3.org/2001/XMLSchema#anyURI";;.
 > >
 > >        urn:oasis:names:tc:xacml:2.0:subject:role
 > >
 > >    RECOMMENDATION: include in next draft.  No objections.
 > >
 > > 60. Define standard "purpose" attributes
 > >
 > >   In order to deal with certain regulatory requirements, it would
 > >   be useful to have a standard "purpose" attribute for resource
 > >   and for action.
 > >
 > >      urn:oasis:names:tc:xacml:2.0:resource:purpose
 > >      urn:oasis:names:tc:xacml:2.0:action:purpose
 > >
 > >   We could also define a standard rule that requires the
 > >   resource purpose to include the action purpose.  This would
 > >   enforce the requirement that data resources only be used for
 > >   the purposes for which they were collected.
 > >
 > >   Issue: purpose should be hierarchical.
 > >
 > >   Policy might have rule saying action-purpose must be "within"
 > >   resource-purpose.  Request might say
 > >   "action-purpose=legal/anti-discrimination/age-discrimination".
 > >   Policy might say "action-purpose within legal".
 > >
 > >   Example: provide shipping address to Amazon.com, but allow only
 > >   for shipping to me.  Insist tagged as resource-purpose
 > >   "marketing/shipping only".  Anyone coming to use this resource
 > >   must have a purpose that
 > >
 > >   Need "Standard Purpose Rule" that will say:
 > >
 > >     action-purpose must be within resource-purpose.
 > >
 > >   i.e. purpose for use must be within the purpose for which it
 > >   was collected.  Purpose for use may be more specific, but must
 > >   not be more general.  This rule would be standard.
 > >
 > >   So should the action-purpose be just the leaf of the hierarchy,
 > >   or should it be the entire hierarchy?  Probably entire
 > >   hierarchy, since leaf names along different branches of the
 > >   hierarchy may conflict.
 > >
 > >   Tim will proceed along these lines in developing a concrete
 > >   proposal.
 > >
 > >   Might want to run this by ISTPA, or get someone there to
 > >   critique the proposal.
 > >
 > >   Should it be in a privacy profile?  Or in 2.0?  For now, assume
 > >   in 2.0 to give people an opportunity to view and critique.
 > >
 > > --
 > > Anne H. Anderson             Email: Anne.Anderson@Sun.COM
 > > Sun Microsystems Laboratories
 > > 1 Network Drive,UBUR02-311     Tel: 781/442-0928
 > > Burlington, MA 01803-0902 USA  Fax: 781/442-1692
 > >
 > >
 > > To unsubscribe from this mailing list (and be removed from the roster of the OASIS TC), go to http://www.oasis-open.org/apps/org/workgroup/xacml/members/leave_workgroup.php.
 > >

-- 
Anne H. Anderson             Email: Anne.Anderson@Sun.COM
Sun Microsystems Laboratories
1 Network Drive,UBUR02-311     Tel: 781/442-0928
Burlington, MA 01803-0902 USA  Fax: 781/442-1692



[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]