[Requests] Comment on Geospatial eXtensible Access Control MarkupLanguage (GeoXACML)

Jan Herrmann jan.herrmann at lmu.de
Tue Jun 19 07:17:59 EDT 2007

Hello every body,


I just figured out that the request for comment period is already over.
Nevertheless I hope that my comments can still be considered.



Part A


1. Evaluator:

Jan Herrmann

Ludwig-Maximilians-University Munich

Luisenstraße 37

80333 Munich



2. Submission: OGC 07-026,

Geospatial extensible Access Control Markup Language (GeoXACML)



Part B



Comment 1: regarding chapter i. Preface:

I agree with Christian Elfers that there is a need to define clearly what is
meant by the expressions “Right Expression Language” and “Policy Language”
if they try to express something different.


Comment 2: changes to the XACML spec that influence GeoXACML

It is necessary to clarify what happens if the XACML spec changes and the
changes influence the GeoXACML spec. 


Comment 3: regarding Chapter Introduction and some general things

It might be helpful to highlight in the introduction, that GeoXACML can be
used to control access to services which can be read, write, update 

Additionally GeoXACML can be used to enforce the policies and rules
respectively on the request to the service (pre-processing) and on the
response of the service (post-processing). 


So far the GeoXACML spec mainly focuses on access control following the
pre-processing mechanism. I would augment and restructure the spec so that
there are two separate chapters at the end, like “enforcing access control
on the request” and “enforcing access control on the response”. In these
chapters it is necessary to explain in detail how an administrator of the
access control system has to define the resource-id element in its
rules/policies depending on whether he/she uses pre- or post-processing
(those of you being able to read german texts can find more information on
these ideas under: http://www.ssg.geo.uni-muenchen.de/geoxacml/da.pdf ) .


Additionally I would describe how GeoXACML has to be used so that the access
control system following the post-processing mechanism implements the
following behaviour:

If there are one or more rules denying access to a subset of the requested
objects than the user should get back the intersection of requested and
accessible features. There are many use cases, where such a behaviour of an
access control system can be very useful. Just imagine you request a very
large amount of data and just one object or attribute is not accessible
(those of you being able to read german texts can find more information on
these ideas under: http://www.ssg.geo.uni-muenchen.de/geoxacml/da.pdf ) .


If some of you are interested I can provide some text modules that describe
the mentioned issues and the consequences for the spec in more detail.



Greets Jan




Dipl. Inf., Dipl.-Geogr. Jan Herrmann

wissenschaftlicher Mitarbeiter

LMU München
Department für Geo- und Umweltwissenschaften
Sektion Geographie
Luisenstraße 37
80333 München

Tel.: +49 (0)89/289 25095
Fax.: +49 (0)89/289 22804
Mobil: +49 (0)163/2509178



-------------- next part --------------
An HTML attachment was scrubbed...
URL: https://mail.opengeospatial.org/mailman/private/requests/attachments/20070619/7a140f31/attachment.htm 

More information about the Requests mailing list