This wiki has undergone a migration to Confluence found Here
August 3rd, 2010 Security Conference Call
Jump to navigation
Jump to search
Contents
Security Working Group Meeting
Attendees
- Tabitha Albertson
- Bill Braithwaite
- Mike Davis Security Co-chair
- Jon Farmer
- Suzanne Gonzales-Webb CBCC Co-chair
- Robert Horn
- Michelle Johnston
- Rob McClure
- Milan Petkovic
- Pat Pyette
- Kenneth Salyards
- Richard Thoreson CBCC Co-chair
- Serafina Versaggi scribe
- Tony Weida
- Craig Winter
Agenda
- (05 min) Roll Call, Call for additional agenda items & Accept Agenda
- PASS Audit Update
- Security and Privacy Ontology project
Minutes
1. Action Items
- Pat Pyette will send a link to the PASS Audit ballot to the Security Work Group list serv with an invitation to submit comments to Draft 0.6. Comments are welcome until the deadline of August 13th.
2. Resolutions - none
3. Updates/Discussion
PASS Audit Update
Pat Pyette provided a brief update on the PASS Audit project
The PASS Audit work has been approved by the SOA Work Group and is ready to go to ballot.
- Little feedback has been received to date. Anyone within the Security Work Group is welcome to submit comments up until August 13th. That will give Pat enough time to incorporate that feedback into the document in time to submit by the deadline for the September ballot which is August 15th.
Security & Privacy Ontology Project
Mike Davis re-re-iterated the focus of the Security and Privacy Ontology
- The OASIS XACML Committee is currently working on a security and privacy project and they are looking to this work group for a beginning ontology in support of that effort.
- Our proposal for approaching this project is to take the RBAC classes from Composite Security and Privacy Information Model, and using an existing HL7 standard – the Permission Catalog, which has a list of objects and actions, create a mini-ontology.
- The draft ontology will be entered into Protégé
- We’ll go through the Information Model class by class, starting with RBAC. When we’re done with all the classes, we’ll be done with the ontology.
- Tony agrees with the approach. He has taken a look at the Permission Catalog from the perspective of creating an ontology from it, but hasn’t looked at it from an XACML perspective.
- In terms of the process, the ontology work will be done off-line and then a straw-man ontology will be brought back to this work group for comments and further refinement.
- Tony walked through the project recap presentation once again for those on today’s call who have not seen it before. The presentation includes links to reference material for background reading on ontologies, OWL 2 and Protégé.
- One of Mike’s comments about the presentation is that it focuses on the technical aspects of creating an ontology which is well and good. But one of the things that Mike is looking for is a narrative description for why we’re creating this ontology
- This is an extension of the Information Model and part of the HL7 SAIF architecture
- Tony: First and foremost, an ontology provides a terminology and vocabulary for all the things of interest. By working together to construct this artifact, we’ll be able to arrive at agreement on exactly what things we’re working on and what names we’re going to use to refer to them.
- By modeling them in OWL and describing their interrelationships and various properties, we can be more precise about what we mean when we use these names. Having done that, this ontology can be the basis for systems which make access control decisions.
- OWL and Protégé themselves can be used to make access control decisions, though this is not a requirement to implement this at run-time. Other technologies can be used at run-time using the ontology.
- We can create various privacy policies that are understandable and exchangeable because they use a common terminology; they can be used to create Privacy Consent Directives which allow individuals to express preferences regardless of the interface people use to express those preferences.
- Mike is looking for a simpler explanation, such as the ontology will make it simpler for individuals to express their preferences.
- The basic policies can be simpler and at run-time, the policy consults the ontologies for specific information that the policy needs, but otherwise you would have to write a policy specifically for each permission where with the ontology you can write a general policy and the ontology is consulted and the specific objects are added at run time.
- In addition, this allows systems to be more adaptive because you update the ontology without re-writing all the business rules.
- There have been some demonstrations in Security using genomic information (the GWAS database), to make access control decisions.
- GWAS is maintained by clinicians and the security system uses the GWAS to identify certain genomic markers that are associated with a specific condition.
- If a new genomic marker is added to the GWAS, the security system consults the GWAS at run-time and adapts accordingly. This is representative of what we’re trying to achieve.
- So we’re creating this ontology for the purposes of simplicity and adaptability
- As we continued through the presentation, additional questions/comments were raised:
- What is an individual?
- An individual is an instance of a class
- Naming convention is to prefix the name with an underscore, e.g.,
- _PodunkGrant_1
- _DoctorWelby
- _NewLabOrder_3
- Naming convention is to prefix the name with an underscore, e.g.,
- An individual is an instance of a class
- By definition a Role is a collection of one of more permissions. Is this described in the example within the presentation?
- At this point, a subject in a role is granted a permission. We could (and probably should) also create connections between roles and permissions
- Categorization of the objects into a hierarchy:
- This can be done along multiple dimensions (axes) simultaneously.
- Record versus workflow objects
- Record: includes things that fall under more than one category (e.g., laboratory and orders)
- Workflow: a token or something that flows through a care plan, treatment plan or a protocol?
- Mike suggested that it could be something like Create Order, Sign Order
- The idea is that people, based on their roles, can be granted permission to participate in certain kinds of workflows, and in the course of those workflows may require access to certain kinds of objects
- Inherent nature of the object, e.g., orders, reports
- Specialty department, e.g., radiology, laboratory
- In terms of ways people can participate in this effort:
- Set Requirements and priorities
- Work specifically on the ontology which would require a working knowledge of Protégé and OWL 2
- Participate in review and provide feedback for the work done between Work Group Meetings
- In terms of the process that we will follow for the development of the ontology, it will be similar to the process we’ve followed for the development of the Security and Privacy Domain Analysis Model
- Guidelines for facilitating these discussions include getting specific items on the agenda depending on the progress on the ontology model
- Provide a progress update since last time ontology was presented or demonstrated
- Questions, comments and suggestions and moving on to new initiatives that have been proposed during the discussion
- The meeting concluded with a review of the background and tutorial material provided in links on the final pages of the recap presentation.
- These references are useful in understanding ontologies as well as the tools that we’ll be using to create this ontology (Protégé and OWL2)
- The latest version of Protégé (4.X) supports OWL2 – and OWL2 has very useful capabilities for expressing some of the aspects of security and privacy that we’re interested in that OWL version 1 does not support
- Also included are links to some example ontologies that have been created by other groups for role-based access control
- These references are useful in understanding ontologies as well as the tools that we’ll be using to create this ontology (Protégé and OWL2)
- The Security and Privacy Ontology project will remain as an item on the Security Work Group agenda going forward. Tony is on vacation this week, but will resume work on the project once he returns from his time off
The Meeting was adjourned at 2:00 PM EDT
No significant decisions or motions were made
