Más contenido relacionado
Más de John Domingue (20)
Super ontology stack_review_m36_051
- 2. Review Agenda, Day I
09.00 Briefing meeting
09.30 Welcome
09.40 SUPER Today, SUPER Successes and a SUPER Future
10.45 Technical Deep Dives
- Part I: Ontology Stack and COBRA
- Part II: Reasoner and WSMO Integration
- Part III: SBP Execution & Analysis
© SUPER 18.06.12 John Kayser (PC/SAP) 2
- 3. Why Semantics?
Reusable
Agent
Business and
Model Activity
sharable
resource
Analyst
Resources
Data
Time
IT specialist
© SUPER 18.06.12 Pierre Grenon, OU 3
- 4. SUPER
Business Process Management
Organisational Context Applications
Business
Process
analyse
execute
BPEL
Reverse Monitoring BPEL4S
Business Behavioural
Analysis sBPEL WS
model Engineering Reasoning
BPMO transform
BPMN support
sBPMN
Semantic Web
transform
Services
sEPC (WSMO)
EPC
© SUPER 18.06.12 Pierre Grenon, OU 4
- 8. Why sBPMN?
■ BPMN is a widely used notation for BPM.
■ SUPER needs an entry point for BPMN
users.
■ BPMN is a graphical notation that needs
explicit semantics (provided by sBPMN).
Lessons learnt:
■ SUPER needs to create entry points.
■ SUPER needs to be independent of
specific notations.
■ SUPER needs to remain up-to-date as
standards change.
© SUPER 18.06.12 Pierre Grenon, OU 8
- 9. sBPMN excerpt
ca. 100 concepts
50+ axioms
© SUPER 18.06.12 Pierre Grenon, OU 9
- 10. Why sEPC?
■ SUPER’s proof of capability to handle
multiple entry points corresponding to
distinct perspectives.
■ EPC is different from BPMN.
■ The sematics for EPC is distinct from
that for BPMN.
© SUPER 18.06.12 Pierre Grenon, OU 10
- 12. Why BPMO?
Intermediate representation for:
■ Abstracting over different notations,
perspectives, levels of understanding
■ Bridging different notations
■ Bridging the world of modelling,
execution and analysis
■ Providing a central platform for
connecting additional knowledge
sources
The above facilitates inter-representational
reasoning and querying.
► Eg, one can pose a query on a BPEL execution trace
and the answer uses knowledge from the original
BPMN diagram and the organisational context.
© SUPER 18.06.12 Pierre Grenon, OU 12
- 13. BPMO excerpt
ca. 70 concepts
© SUPER 18.06.12 Pierre Grenon, OU 13
- 14. Why sBPEL?
■ sBPEL is SUPER’s ontology at the
receiving end of ontology-based model
transformation (mediated by BPMO)
hooking into execution (BPEL through its
BPEL4SWS extension).
■ sBPEL and BPEL4SWS have hooks
allowing for the execution of SWS and of
their compositions.
■ We can reason about sBPEL, that which
is about to be executed and that which
comes after execution (thus supporting
analysis).
© SUPER 18.06.12 Pierre Grenon, OU 14
- 16. Why BRO?
■ Support high level reasoning about
behaviour
► E.g. compliance checking, composition validation
■ This reasoning can include any
knowledge carried within the ontology
stack.
© SUPER 18.06.12 Pierre Grenon, OU 16
- 17. BRO
20+ concepts
3 instances
ca. 20 axioms
(BPMO process)
© SUPER 18.06.12 Pierre Grenon, OU 17
- 18. Why UPO?
■ Umbrella ontology for the stack
■ Gathers common concepts and
relations
■ Facilitates inter-ontology reasoning
50+ concepts
20+ instances
ca. 20 relations
3 relation instances
20+ axioms
© SUPER 18.06.12 Pierre Grenon, OU 18
- 19. UPO’s usage is embedded in
that of BPMO
BPMO
BPMO UPO
© SUPER 18.06.12 Pierre Grenon, OU 19
- 20. Why a domain ontology stack?
■ Resources for enriching the
representation of BP models based on
BPMO.
■ Domain ontologies provide values for
attributes of BPMO process instances.
■ Vertical domain with three levels of
generality:
► Generic business domain (organisation) ontologies
► Extended by telecommunication ontologies
► Extended by use case specific ontologies.
© SUPER 18.06.12 Pierre Grenon, OU 20
- 21. UPO is extended by
Organisation Ontologies
BFO BMO
BPRO
OSO
OUO
BRONTO
ResOnt
© SUPER 18.06.12 Pierre Grenon, OU 21
- 23. Organisation Ontologies excerpt (2)
BRONTO ResOnt
20+ concepts 3 instances 50+ concepts 30+ instances
4 relations ca. 20 axioms 10+ relations 3 relation instances
30+ axioms
© SUPER 18.06.12 Pierre Grenon, OU 23
- 25. Organisation Ontologies excerpt (4)
16 concepts 36 concepts
BPRO
deo dpo
dro ecao
25 concepts 10 concepts
9 instances
© SUPER 18.06.12 Pierre Grenon, OU 25
- 27. Why NGOSS Ontologies
(YATOSP)?
■ Support communication and inter-
operability in the telecommunication
domain
■ Provide an easy entry point for using
SUPER in the telecommunication
domain
■ NGOSS ontologies correspond to a set
of industry standards
© SUPER 18.06.12 Pierre Grenon, OU 27
- 28. NGOSS excerpt
Telco business IT systems
eTOM process functions TAM for process execution
ca. 370 concepts 200+ concepts
360 instances 1K+ instances
4 axioms 5 axioms
© SUPER 18.06.12 Pierre Grenon, OU 28
- 29. NGOSS excerpt (2)
Common
SID domain NBC
vocabulary
Telco business
process contract
ca. 1.3K concepts
ca. 1.2K instances 14 concepts
4 relation instances ca. 220 instances
8 axioms 3 relations
© SUPER 18.06.12 Pierre Grenon, OU 29
- 31. Why use case ontologies and
knowledge bases?
■ Each enterprise has its specifics:
► Organisational structure
► Variants and additions to BP models
► Data structure.
■ Each enterprise has its own data.
■ Common umbrella framework and
common representation formalism.
© SUPER 18.06.12 Pierre Grenon, OU 31
- 33. Example of Application
(from TID Use Case)
Ontology Extension Process Instance
© SUPER 18.06.12 Pierre Grenon, OU 33
- 36. SUPER Process
Ontology Stack
© SUPER 18.06.12 Pierre Grenon, OU 36
- 38. UPO excerpt
50+ concepts
20+ instances
ca. 20 relations
3 relation instances
20+ axioms
© SUPER 18.06.12 Pierre Grenon, OU 38
- 39. BRONTO excerpt
20+ concepts
3 instances
4 relations
ca. 20 axioms
© SUPER 18.06.12 Pierre Grenon, OU 39
- 40. OSO & OUO excerpt
9 concepts
11 relations
19 concepts
© SUPER 18.06.12 Pierre Grenon, OU 40
- 41. ResO excerpt
50+ concepts 30+ instances
10+ relations 3 relation instances
30+ axioms
© SUPER 18.06.12 Pierre Grenon, OU 41
- 42. ResO excerpt
50+ concepts 30+ instances
10+ relations 3 relation instances
30+ axioms
© SUPER 18.06.12 Pierre Grenon, OU 42
- 44. eTOM
ca. 370 concepts
360 instances
4 axioms
Telco business
process functions
© SUPER 18.06.12 Pierre Grenon, OU 44
- 46. NBC
14 concepts
ca. 220 instances
3 relations
Telco business
process contract
© SUPER 18.06.12 Pierre Grenon, OU 46
- 47. TAM
IT systems
for process execution
200+ concepts
1K+ instances
5 axioms
© SUPER 18.06.12 Pierre Grenon, OU 47
- 49. sBPMN usage in SUPER
tools and components
SUPER Execution SUPER Tooling
Semantic
Semantic Monitoring &
BPEL Modelling
Execution Execution Tool Management Analysis Tool
Environment Tool
Engine
Semantic Service Bus Deployment
Event Sink
XPDL2sBPMN
Protocol Binder
and
SBP SBP Process
Composition Mediation
Business
Semantic Web
Services
Execution sBPMN2BPMO
Process Library History
SBP Discovery Data Mediation Repository
translators
SUPER Repositories
SBP Reasoner Transformation
SUPER Plafform Services
© SUPER 18.06.12 Pierre Grenon, OU 49
- 50. sEPC usage in SUPER
tools and components
SUPER Execution SUPER Tooling
Semantic
Semantic Monitoring &
BPEL Modelling
Execution Execution Tool Management Analysis Tool
Environment Tool
Engine
Semantic Service Bus Deployment
Event Sink
Protocol Binder
SBP SBP Process
Composition Mediation
Semantic Web
Business Execution
Services
SBP Discovery Data Mediation
Process Library
Repository
History
sEPC2BPMO
SUPER Repositories translator
SBP Reasoner Transformation
SUPER Plafform Services
© SUPER 18.06.12 Pierre Grenon, OU 50
- 51. BPMO usage in SUPER
tools and components
SUPER Execution SUPER Tooling
Semantic
Semantic Monitoring &
BPEL Modelling
Execution Management Analysis Tool
Execution
Engine
Environment
Tool
Tool BPMO Modeller
Semantic Service Bus Deployment BP Mediator design-time
Event Sink
Translators:
Protocol Binder
BPMO2SBPEL
sEPC2BPMO
SBP SBP Process
Composition Mediation sBPMN2BPMO
Semantic Web
Business Execution
Services
Process Library History
SBP Discovery Data Mediation Repository
BPMO API
SUPER Repositories
SBP Reasoner Transformation
SUPER Plafform Services
© SUPER 18.06.12 Pierre Grenon, OU 51
- 52. sBPEL usage in SUPER
tools and components
SUPER Execution SUPER Tooling
Semantic
Semantic Monitoring &
BPEL Modelling
Execution Management Analysis Tool
Execution
Engine
Environment
Tool
Tool BPMO2sBPEL
Semantic Service Bus Deployment and
Event Sink
sBPEL2BPEL4SWS
Protocol Binder
translators
SBP SBP Process
Composition Mediation
Semantic Web
Business Execution
Services
Process Library History
SBP Discovery Data Mediation Repository
SUPER Repositories
SBP Reasoner Transformation
SUPER Plafform Services
© SUPER 18.06.12 Pierre Grenon, OU 52
- 53. BRO usage in SUPER
tools and components
SUPER Execution SUPER Tooling
Semantic
Semantic Monitoring &
BPEL Modelling
Execution Management Analysis Tool
Execution
Engine
Environment
Tool
Tool BP Mediator Reasoning
Semantic Service Bus Deployment Support validation of
Event Sink
composition
Protocol Binder
BPMO2BRO translator
SBP SBP Process
Composition Mediation
Semantic Web
Business Execution
Services
Process Library History
SBP Discovery Data Mediation Repository
SUPER Repositories
SBP Reasoner Transformation
SUPER Plafform Services
© SUPER 18.06.12 Pierre Grenon, OU 53
- 54. RBEO usage in SUPER
tools and components
SUPER Execution SUPER Tooling
Semantic
Semantic Monitoring &
BPEL Modelling
Execution Management Analysis Tool
Execution
Engine
Environment
Tool
Tool Semantic Business
Process Discovery
Semantic Service Bus Deployment Engine
Event Sink
(sRBE tool)
Protocol Binder
SBP SBP Process
Composition Mediation
Semantic Web
Business Execution
Services
Process Library History
SBP Discovery Data Mediation Repository
SUPER Repositories
SBP Reasoner Transformation
SUPER Plafform Services
© SUPER 18.06.12 Pierre Grenon, OU 54
- 56. Why NGOSS Ontologies
(YATOSP)?
■ The Telemanagement Forum (TMF) is an
association of telecommunication enterprises and
IT providers.
■ The TMF’s development of the New Generation
Operation Support Systems (NGOSS) results,
among other things, in a set of standards.
SID Shared Information Domain Common domain vocabulary
NBC NGOSS Business Contract Telco business process contract
eTOM Enhanced Telco Operation Map Telco business process functions
TAM Telco Applications Map IT systems for process execution
■ These were ontologised in SUPER.
© SUPER 18.06.12 Pierre Grenon, OU 56
- 57. Why an ontology stack?
■ A stack is formed of elements which are
putatively independent to some degree.
■ The degree of independence depends on
application context sampling the stack
(facilitating partial reuse).
■ Also in some cases, size matters and the break
down of ontological resources can be motivated
by efficiency and technical parameters.
■ Finally, SUPER is a collaborative project with
distributed competences and responsibilities.
© SUPER 18.06.12 Pierre Grenon, OU 57
- 58. Why ontologies?
■ Representing reality
■ Bridging between different representations, possibly
of the same things using different representation
formalisms
■ Providing explicit semantics to existing
representation formalisms
■ Abstracting from existing representations
■ Supporting updatable and potentially growing
interoperability with centralised mapping
■ Supporting scalability through mechanisation and
automatised translation
■ Supporting sharing and reuse over the Semantic Web
© SUPER 18.06.12 Pierre Grenon, OU 58
- 59. Why Ontology-based
Representation for BPM?
■ Heterogeneous sources and viewpoints abstracted
into a single ontology-based representation of
processes, their structures, relations, and
properties for which BPMO provides the backbone.
► Translation from and/or to BPMN, sEPC, BPEL
■ Heterogeneous formalisms streamlined into a
single knowledge representation formalism
(WSML):
► bringing BP Models to machines
► facilitating integration and interoperability
► allowing automatisation, execution, reuse
► in an accessible and scalable manner.
© SUPER 18.06.12 Pierre Grenon, OU 59