"Seegrid will be due for a migration to confluence on the 1st of August. Any update on or after the 1st of August will NOT be migrated"

Workshop on Register-managed vocabularies in a service-oriented architecture

Contents

Related pages



Outcomes

Thanks for all contributing to make this workshop productive, informative and ultimately successful in finding a way forward with the vexing questions of vocabulary management.

A (very quick) summation of the key lessons, pending a fuller report...

  • The business requirements for a "Register" were explored, and proved to be poorly supported by the current registry interfaces standards. Minor changes to support convenient management of registries were deemed worthy for consideration in the upcoming ebXML RegRep v 4.0 specification revision. * There is a need for a heterogeneous register environment, and thus for the ability to locally proxy (eg cache) remote registers. This provides for an abstraction between the physical form and location of key vocabulary resources and the form we need them locally. This in turn allows us to charge ahead and build infrastructures safely if the "remoting" model is formalised so that local implementations maintain integrity if the source of the content moves towards better managed resources in the future (eg. from PDFs to registry services)
  • A number of infrastructure building activities got to meet and greet and share common visions. Ongoing collaborations are expected to result.
  • A technical implementation approach was chosen for the task of creating and exercising formalisms for service profile descriptions and query invocations.

Follow up activities

A Use Case scenario was developed

Ongoing work aims to develop a common (reusable) architectural package of Use Cases for cross-system vocabulary sharing using registry services.

The priority areas to be elucidated are:
  • basic Register concept, linking Use Cases to ISO 19135 concepts
  • use of a remotely defined vocabulary in a system, using a register concept
  • metadata management using controlled vocabularies and profiles - CSIRO will be exploring the use of [ OASIS CAM | http://www.oasis-open.org/committees/tc_home.php?wg_abbrev=cam ] (Content assembly mechanism) in upcoming projects as part of developing an approach to creating and maintain Data Product Specifications.

  • Packaging of Use Case models:
    packages.PNG

Initial drafts of Use Cases:

  • SharedRegister.xml: Draft Use Case model for register proxying (Enterprise Architect XMI)

 

Context

Summary of Meeting

(Farrukh Najmi)

Special thanks to Rob and Simon for organizing the logistics, agenda and cross-organizational participants so very well. It made for a very productive and creative week of collaboration.

An important take-away from our collaboration was the valuable feedback on the OASIS ebXML RegRep standard. Our discussion on the use of federated registers to support model-driven authoring of service profiles and data product specifications is particularly interesting as it provides a flyweight pattern for efficiently managing governance policies over related set of metadata.

Our discussion on local caching of remote registers governed by external organizations is central to enabling reuse of information such as standard vocabularies and ontologies across organizational boundaries which in turn is required for interoperable services.

The current ebXML RegRep 3 standard is quite flexible and extensible and can meet many of the foundational requirements to support the use cases you have articulated. However, it is lacking the concept of registers to fully support the requirements for independent and distributed governance of metadata collections by different organizations.

ebXML RegRep 3 standard provides a powerful federation capability to support distributed governance. However, this capability is complex to use in practice. Again, our discussion on simplified local caching of registers through the use of an enhanced subscription model seems promising to address the current complexity in ebXML RegRep 3 federation model.

Our face-to-face discussions with various groups has been vital to mutual sharing of ideas and has led to our common understanding on the use cases and requirements for distributed governance and reuse of standard vocabularies in support of model-driven authoring of service profiles and data product specifications. It has also identified some important gaps in current ebXML RegRep standard and we have discussed potential ways to address these gaps.

Summary of Shared Vision

(Farrukh Najmi)

The ultimate vision I took back from the workshop is one where organizations like CSIRO and its partners in a broader "cyberinfrastructure" for research are able to stand up registers containing their data and metadata with relatively low effort and cost, and integrate with other infrastructures as they emerge. Potentially, a central registry provides coordination and identifier resolution capability across the distributed registries containing all participating registers. Each community might establish a coordinating node for its infrastructure, and share information with other communities as they emerge. This model will be greatly facilitated by commonality of the meta-model for such registers, and implementation standards, such as provided by OASIS.

The information in these registers may used to support composition of metadata views from reusable pieces, for example by authoring tools to create data product specifications and service profiles. New profiles may extend existing profiles by adding restrictions defined by their specific deployment needs. New service instances are easily validated for conformance to the service profiles they claim to support. Most importantly, each partner organization manages their own registers using their own governance policies.

From our discussions it would appear that the vision proposed is quite feasible and desirable, but would take a little effort to articulate, test and promote through the standardization process.

I hope that we continue this collaboration to the logical next step which is to identify and implement a pilot project.

 

Original workshop Plan

Purpose

What it is

The purpose of the workshop is to establish a baseline of understanding how to access content held in registers as vocabularies in the many places (particularly XML artefacts) where they need to be referred to.

It is a coming together of practitioners whose project environments and technical concerns overlap. Because any solution will rely on the governance of infrastructures, this is intended to start the process of aligning technical solutions and governance realities.

The idea is to establish openness in the process, whilst staying focussed on having a short term implementation path identified. An expert consultant, Farrukh Najmi, will be brought in to review the requirements and provide insight into the overlap with OASIS and W3C architectures, and facilitate the development of an approach.

Open forums will be held to describe preliminary scope and thinking, in order to aid the development of more formal follow-up liaison and alignment activities.

What it is not

This is not intended to be the presentation of a solution to the broader community, nor a wide-ranging requirements gathering or validation exercise.

It is not the one an only opportunity to influence the development of such solutions, it is merely an alignment of certain infrastructure development requirements with the emerging thinking in the international standards-setting bodies.

so, dont fret if you are interested, or the results apply to your domain - there will be follow-up and open discussion opportunities.

Key Outcomes desired

  • recommended approach for immediate technical requirements (Web Service Profile, data product specification and data model profiles)
  • establishment of mutual visibility between our investigations and key international standards effords
  • informing discussion regarding governance of key registers and standards adoption processes within local (Australian) context, both government and academic sectors

Agenda

Monday 28 April. Initial sanity check

Project oriented technical briefings: explaining positions: what WRON-RM, AusScope etc wants to address and why, what ebXML and semantic web world are currently doing Venue: Novotel Wollongong Participants: Rob Atkinson, Simon Cox, Farrukh Najmi

Tuesday 29 April

Discuss broader practice, governance arrangements and identify key discussion points for Wed forum.

In brief, what messages do we need to pass to, or receive from, the governance of e-infrastructures, based on the technical analysis to date,

Venue: Canberra, CSIRO Christian Labs Boardroom, and video conferencing facility Time: 11am to 5pm Participants: NZ State Services Commission via Videolink. Open Invitation

Wednesday 30th April Public forum

broader community involvement in workshop on enablement of pan-institutional interoperability through registry infrastructure:
  • WRON/NCRIS introductions - why we are greasing the wheels of alignment
  • OSDM and AGIMO perspective
  • ABS perspective
  • WRON-RM analysis of requirements
  • Industry perspective - Farrukh Najmi
  • summary of NZ government experience
  • half day facilitated discussion to identify priorities, roles, mandates

Venue: Time: 10am - 4pm Venue: Burley Griffin Room, Griffin Hotel, 15 Tench Street, Kingston

Lunch and morning/afternoon tea will be provided.

Participants: OSDM invitees

Thursday May 1 Technical Discussions

exploring the problem of system-level interoperability and architecture design - modularity, inheritance and profiling

Proposal: review of "Master Data Management" concept: http://msdn2.microsoft.com/en-us/library/bb190163.aspx#mdm04_topic4

Venue: CSIRO IM&T, Alan Brown Conference room Building 1 level 1, Banks Street, Yarralumla ACT 2601 (parking free onsite)

Friday May 2 Developing recommendations for Follow-up activities

presenting thinking to internal project groups, discussing roles of standards organisations, and ways forward through collaborations, and addressing specific issues not covered in time.

 

Topic attachments
I Attachment Action Size Date Who Comment
RegisterProxy.rtfrtf RegisterProxy.rtf manage 676.1 K 03 Jun 2008 - 14:25 RobAtkinson Generated documentation view of Use Case model for Registers and metadata composition
SharedRegister.xmlxml SharedRegister.xml manage 431.1 K 03 Jun 2008 - 14:27 RobAtkinson Draft Use Case model (Enterprise Architect XMI)
VocabUsageScenario_(2).pdfpdf VocabUsageScenario_(2).pdf manage 302.6 K 15 May 2008 - 12:03 RobAtkinson Use Case Scenario for eResearch workflow using registered vocabularies
packages.PNGPNG packages.PNG manage 2.6 K 04 Jun 2008 - 09:55 RobAtkinson Packaging of Use Case models
Topic revision: r14 - 15 Oct 2010, UnknownUser
 

Current license: All material on this collaboration platform is licensed under a Creative Commons Attribution 3.0 Australia Licence (CC BY 3.0).