GOLD

1
GOLD

• an infrastructural approach to virtual
  organisations

2
about the project

• Lots of VO definitions
• All different
• None entirely satisfactory
• Lots of different types of VO
• Greatly differing requirements
• Different degress/aspects of virtual
• Hard to write a definition that includes them all
  without being too general

3
types of VO

• Very different types of organisation that involve
  some aspect of virtual working
• Franchises, such as fast food outlets, building
  society counters, travel agents etc.
• Utilities companies such as gas, electric,
  railway etc. selling services based on a
  separately-managed infrastructure. Also, the
  company or companies managing that
  infrastructure.
• Companies with long-term outsourcing agreements.
  For example, outsourcing of IT or payroll
  services.
• Companies using short-term outsourcing agreements
  to meet particular needs on-demand. For example,
  chemicals companies requiring specific expertise
  scientific companies needing additional compute
  resource supply-chain based companies wishing to
  respond flexibly to exceptions etc.
• Truly distributed companies with different
  functions supplied by different organisations but
  no central management function.

4
infrastructural approach

• Diverse requirements demand an infrastructural
  approach
• Core services implementing VO behaviour
• VOs at or close to application level
• Maximum flexibility
• Loose coupling
• Needed
• Standards
• Middleware
• Infrastructure
• Languages

5
VO infrastructure

• Coordination
• Information management
• Security
• Trust

6
proof of concept

• The infrastructure will be tested by applying it
  to RD projects within the chemicals industry
• ProjectVO
• Highly dynamic
• Membership
• Focus
• Basis of interaction between parties

7
progress requirements

• Conventional approaches are difficult when
• The technology doesnt exist
• The business use doesnt exist
• The potential users have no basis for thinking
  about requirements
• They dont know what they want to do
• They dont know what their non-functional
  requirements are (security, dependability etc.)
• An ethnographic approach is being taken
• Interviews with target organisations
• Studies of how they do things at the moment
• Understanding of business context
• Iterative development in step with requirements
  capture

8
progress requirements

• Interviews with several chemicals companies have
  taken place
• These have confirmed some of our ideas about VOs
• And have changed others
• A great deal of enthusiasm
• A diverse set of functional and non-functional
  requirements is being gathered
• We are beginning to understand the real business
  context what VOs can do, how people want to use
  them, what is their potential? (more on this
  later)
• SSM models of organisations constructed

9
progress coordination and trust

• Distributed workflow enactment
• Our own language with semantics defined using pi
  calculus
• Compositions can be model checked using SPIN
• Basic system works, basic GUI implemented
• Further work needed on GUI, composition/deployment
  tools, monitoring, management
• Further work required on defining tasks, task
  management, communication etc.

10
progress coordination and trust

• Contract management
• Contracts mediate interactions between
  organisations, describing rights, obligations,
  constraints etc.
• This provides evidence in the case of disputes
• And can be used for coordination
• Promela used as a specification language for
  describing contracts
• Basic contract mediators implemented
• Need to investigate how to integrate with
  workflow enactment and access control

11
progress coordination and trust

• Non-repudiable interactions
• Actions must be attributable to the parties who
  performed them and commitments to the committing
  party
• Non-repudiable interactions implemented on J2EE
• Works for interactions between Java objects over
  RMI
• Currently being re-implemented for SOAP using
  appropriate WS standards
• Prototype working

12
progress coordination and trust

• Summary
• We can run business processes over organisational
  boundaries
• We can specify and police interactions using
  electronic contracts
• We can construct non-repudiable audit trails
• Integration needed

13
progress information management

• Each party has heterogeneous, distributed and
  autonomous information sources
• These need to be integrated in non-trivial ways
• Dynamic information integration/sharing and
  business composition
• Dynamic security
• Notification
• Ontology
• Discovery
• Personalisation

14
progress information management

• Test/development platform built (based on Oracle)
• Requirements captured
• Architecture proposed
• Services identified (IAS, DBPMS, LCS etc.)
• Significant interest from Oracle
• Need to
• define services
• Formalise the information model
• Implement

15
progress security

• Looking for a formal answer to the question what
  is a dynamic coalition?
• Models of
• What a VO is
• What knowledge is
• How it is transferred
• Some models produced and will be used to analyse
  information flows in VOs

16
progress security

• RBAC and TBAC have been investigated
• Dynamic AC investigated
• Can the infrastructure dynamically control access
  based on policies?
• Prototype constructed based on existing
  technologies

17
progress general

• Candidate architecture proposed
• Core services defined
• Framework for building and managing VOs
  constructed
• Demos available (at this meeting) in all major
  areas

18
summary

• The project has been running for about 1 year
• Recruitment is complete (visa problems
  notwithstanding)
• Progress has been very good
• Strong understanding of industrial requirements
• Candidate architecture
• Prototypes in most major areas
• Lots of talks/outreach events
• Still a great deal to do!
• Architecture
• Integration
• Testing