I'm for a "meta model", represented in the spec as a required "Access
Policy Module", with a non normative XACML binding.
-Matt
On Wednesday, January 8, 2003, at 12:37 PM, Damodaran, Suresh wrote:
> Here is a thought that we may need to confront sooner or later.
> XACML is not free of IP claims [1,2]. Since we don't want reg-rep v3.0
> to be encumbered by IP claims, one option we have is the following:
>
> 1. Make a meta model and then bind XACML to it. This should leave the
> option
> of making other
> bindings as well. I don't claim I know exactly how to do this as yet,
> but that is something we would need to figure out together.
>
> The second option is to forget about the metamodel and let the burden
> fall
> on
> the implementers.
>
> Any other thoughts or other options?
>
> In any case, it looks prudent to cleanly identify and compartmentalize
> the spec portions that deal with XACML and Custom Access Control.
>
> Regards,
>
> -Suresh
> Sterling Commerce (on loan to RosettaNet)
> [1] http://www.oasis-open.org/committees/xacml/ibm_ipr_statement.shtml
> [2] http://www.oasis-open.org/committees/xacml/cg_ipr_statement.shtml
>
>
>
> -----Original Message-----
> From: Farrukh Najmi [mailto:]
> Sent: Tuesday, January 07, 2003 8:03 PM
> To: Breininger, Kathryn R
> Cc: Matthew MacKenzie; Damodaran, Suresh;
> Subject: Re: [regrep] XACML and Access Control Policy
>
>
> I believe that proposed changes for custom ACP are largely orthogonal
> to
> the the set of changes proposed to be reviewed this Thursday. The only
> overlap in in the security chapters of RS and RIM where the changes for
> 2.33 were fairly minor. We could defer these chapters review until we
> finish the Custom ACP task.
>
> --
> Regards,
> Farrukh
>
>
>
> Breininger, Kathryn R wrote:
>
>> Sounds like this should be the first agenda item. Do you anticipate
>> other
> sections of the specs changing as a result? If the second agenda item
> is
> reviewing the current changes, are there sections that will be
> affected by
> this proposal that we should skip in our spec review?
>>
>>
>> On Monday, Jan 6, 2003, at 17:03 America/Vancouver, Farrukh Najmi
>> wrote:
>>
>>
>>
>>> Suresh,
>>>
>>> XACML based custom access control policy was planned for V3 and is in
>>> fact the only task that was planned for V3 that we have not addressed
>>> for V3. The task was dropped for two reasons:
>>>
>>> -XACML was a moving target
>>>
>>> -We had no one signed up for the task
>>>
>>> Given that XACML is now a month away from becoming the next OASIS
>>> approved standard ( I believe it will get approved) and given that
>>> you
>>> are offering to take ownership of the, I completely agree with your
>>> suggestion that we should do it for V3.
>>>
>>> My experience with several strategic ebXML Registry pilots using the
>>> ebxmlrr project has shown that this is a *MUST* feature for V3. In
>>> fact the ebxmlrr project has been implementing XACML based custom ACP
>>> as a implementation specific feature already. The experience further
>>> suggests that XACML is ready for building our specs on top of and
>>> that
>>> we *SHOULD* do custom ACP for V3 based on XACML.
>>>
>>> I believe we could accommodate the increase in scope with about 1
>>> month slip to our V3 schedule. I think that the benefit of having
>>> this
>>> strategic feature far outweighs the cost of the delay to V3 schedule.
>>>
>>> I would be very willing to help you with this task. Maybe Sanjay
>>> could
>>> help as well (Sanjay?) and we could get our security sub-team charged
>>> up for V3.
>>>
>>> Kathryn, I propose we add this issue to this week's TC con-call.
>>>
>>> --
>>> Regards,
>>> Farrukh
>>>
>>>
>>>
>>> Damodaran, Suresh wrote:
>>>
>>>
>>>
>>>> Hi all,
>>>>
>>>> It would be great to have XACML based custom access control policy
>>>> for V3. Is this something we are considering for V3?
>>>>
>>>> I may even volunteer sometime:-)
>>>>
>>>> Best regards,
>>>>
>>>> -Suresh
>>>> Sterling Commerce (on loan to RosettaNet)
>>>>
>>>>
>>>>
>
>
> ----------------------------------------------------------------
> To subscribe or unsubscribe from this elist use the subscription
> manager: <http://lists.oasis-open.org/ob/adm.pl>