NCOICBimsonSirmons08062006
Download
Report
Transcript NCOICBimsonSirmons08062006
Case Study: Deploying an Ontology-Based
Enterprise Information Integration Solution
at the USAF 45th Space Wing
Dr. Kent D. Bimson, Bimson Consulting
Mr. Richard Sirmons, USAF 45 RMS/RMRC
Network Centric Operations Industry Consortium*
August 1, 2006
*Based on Presentation at
Semantic Technology Conference
March 7, 2006
STC 2006
Bimson & Sirmons
1
Contact Information
Mark Wallace, Modus Operandi
[email protected]
Alan McCutchen, Modus Operandi
[email protected]
Dr. Perakath Benjamin, KBSI
[email protected]
Dr. Renato Levy, IAI and Dr. Lenny Haynes, President, IAI
[email protected]
[email protected]
Richard Thiebauth, USAF 45RMS/RMRC
[email protected]
Gary Tupper, Scitor
[email protected]
Dr. Kent Bimson, Bimson Consulting
[email protected] 321 223-9751
STC 2006
Bimson & Sirmons
2
Contents
Problem
Vision
Strategy
Approach
Funding
Architecture
Applications
Ontology Development Process Overview
Ontology Life Cycle Management Overview
Way Ahead
STC 2006
Bimson & Sirmons
3
Problem:
The Fragmented Enterprise
STC 2006
Bimson & Sirmons
4
Problem: Information Sharing Today
KSC
Florida Space
Authority
DELTA IV
30 SW
Navy
Down range
sites
Commercial
Users
Delta II
Other
Agencies
RTSC
45 SW
LOSC
JBOSC
SMC/RNP
HQ AFSPC
SLRSC
STC 2006
14 AF
ATLAS V
CSSC
Bimson & Sirmons
Spacecraft
users
5
Result: Fragmented Range Picture
Many obstacles to information access, integration and sharing
Different owners and agendas
Incompatible formats
Different identifiers
STC 2006
Stored in many locations
Structured and unstructured
Different naming conventions
Bimson & Sirmons
6
Vision:
Single Integrated Range Picture
STC 2006
Bimson & Sirmons
7
Vision
The Five Rights
Integrating and sharing information from multiple sources
to deliver the right information to the right person at the
right time in the right form to make the right decision
Single Integrated
Range Picture
STC 2006
Bimson & Sirmons
8
Shrink-Wrapped Solutions Do Not Exist!
STC 2006
Bimson & Sirmons
9
Strategy and Team:
Knowledge Management Initiative
STC 2006
Bimson & Sirmons
10
Strategy: Three Legs of KMI
Knowledge Management Framework
Technical integration infrastructure
Knowledge Management
Initiative
STC 2006
Bimson & Sirmons
11
Strategy Description
1. COI: gaining access to data and agreement on term definitions
Community of Interest for Space Launch & Range Operations
Teamwork in technology, application development, deployment
2. KMF: building infrastructure for data integration & sharing
Integrates distributed structured and unstructured information
using an ontology, based on MO Wave and BEA AquaLogic DS
In use at the USAF 45th Space Wing
Provides integrated data to new/existing applications to form Single
Integrated Range Picture (SIRP)
Being extended to support Ontology Life Cycle Management in the
Ranger SBIR (Modus Operandi) and ODIF SBIR (KBSI)
3. SIRP: creating Single Integrated Range Picture applications
Applications that deliver KMF-integrated data to diverse users
Use the KMF to access unified structured/unstructured information
STC 2006
Bimson & Sirmons
12
Technical Team: Roles and Responsibilities
KMI/KMF and SIRP Apps, funding
SBIR funding
SBIR funding
KMF (Wave & BEA) ontology-based
EII Ontology life cycle mgmt
Shark C2 System
Knowledge Based Systems, Inc. Ontology-driven search
Text mining for concepts/relations
Intelligent Automation, Inc.
Ontology-based scheduling engines
Ontology-driven agent-based system
Scitor, Inc.
Management & ontology support
Space Coast Launch Services
Deployment, applications, admin
Computer Sciences Raytheon
Applications
Bimson Consulting
KMI strategy, SBIR consulting
Ontology development and training
USAF 45th Space Wing
Missile Defense Agency
Air Force Research Lab
Modus Operandi, Inc.
STC 2006
Bimson & Sirmons
13
Technical Approach:
Knowledge Management Framework
& SIRP Applications
STC 2006
Bimson & Sirmons
14
KMF and SIRP Appllications
Launch
Facilities
Database
KMF: Accesses Related Data
Weather
Alerts
Vehicle
Processin
g
Database
KMF: Integrates Related Data
KMF
SIRP Applications: Present Integrated Data
Airfield
Status
Schedule
Database
Knowledge Management Framework and SIRP Applications
provide a Single Integrated Range Picture
STC 2006
Bimson & Sirmons
15
Building KMF and Applications
2 MDA SBIRs building KMF
for unstructured data
Complete: Jun 07
Applications: USAF 45SW, AFRL and MDA
Delivery Feb 06 to May 07
Shark Command and Control System (SC2S)
Airfield Operations support
Range Instrumentation Status Display
KMF
Range system SITREP generation
ODIF MDA SBIR: Ontology-driven text search
Star Gate AFRL SBIR: Ontology-driven schedule
integration and schedule engine generation
2 AFRL SBIRs building
KMF for structured data
Complete: Apr 06
STC 2006
Wave search and browse tools
Wave ontology LCM tools: Ranger MDA SBIR
Bimson & Sirmons
16
SIRP Application Description
Example domain-specific applications include:
Shark Command and Control System (SC2S): Wing Operations
Center application for battlestaff using WebEOC integrated with
KMF (Modus Operandi and SCLS-early deployment)
Range Instrumentation Status Display: status of instrumentation
during launch countdown (CSR-early deployment)
StarGate SBIR: agent-based, ontology-driven schedule generation
engine (IAI-in development)
Example general applications include:
ODIF SBIR: ontology-directed search and concept mining (KBSI-in
development)
Wave Search: search through structured and unstructured data
(MO-early deployment)
Wave View: springy-graph ontology browsing of integrated data
(MO-early deployment)
Ranger: ontology life cycle management tools (MO-in development)
STC 2006
Bimson & Sirmons
17
Funding:
Leveraging SBIR Programs
STC 2006
Bimson & Sirmons
18
KMI Funding: Leveraging SBIR Programs
2 MDA SBIRs building KMF
for unstructured data
Complete: Jun 07
Applications: USAF 45SW, AFRL and MDA
Shark Command and Control System (SC2S)
Airfield Operations support: Jan 06
Range Instrumentation Status Display
KMF
Range system SITREP generation: Jan 06
ODIF MDA SBIR: Ontology-driven text search
StarGate AFRL SBIR: Ontology-driven schedule
integration and schedule engine generation
2 AFRL SBIRs building
KMF for structured data
Complete: Apr 06
STC 2006
Wave search and browse tools: Mar 06
Wave ontology LCM tools: Ranger MDA SBIR
Bimson & Sirmons
19
KMF Architecture:
Wave Ontology-Based EII Platform
STC 2006
Bimson & Sirmons
20
KMF Architecture: MO Wave & BEA ALDS
Applications
Shark C2 System
ODIF
StarGate
Wave View & Search
Ontology Life-Cycle Management
© 2006 Modus Operandi, Inc.
Used by permission
…
Ontology-Driven Web Services
*** Ranger Phase II ***
BEA AquaLogic Data Services Platform
Structured / Unstructured Data ►
DB
DB
Wave Unstructured Data
Index, Search, Access
BEA WebLogic Server
STC 2006
OWL DL Reasoners
DB
XML
Bimson & Sirmons
XML
21
SIRP Applications:
Analyzing Integrated Information
STC 2006
Bimson & Sirmons
23
Shark C2 System, SITREP and RISD
Range Instrumentation
Status Display and
Situation Report
(SITREP) generation
STC 2006
SC2S leverages COTS
product (WebEOC) as basis
for improved information
support to WOC
Bimson & Sirmons
24
ODIF Ontology Search & Concept Mining
Protégé ontology used to drive ODIF search
STC 2006
Bimson & Sirmons
25
StarGate Agent-Based Scheduling
configurable Gant
chart area
select resources by
type
quick info
list of resources
detail information
about a resource
allocation
STC 2006
Bimson & Sirmons
26
StarGate: Querying for an Operation
Enter the Operation Number
STC 2006
Bimson & Sirmons
27
StarGate: Results
Operation status
Cannot change, but view
Customer/ ODType/
Annex or AOP
Can however change
dates, request additional
resources
Original schedule
History
User can scrub, re-schedule,
or cancel an operation; Or
update to indicate completed
STC 2006
Bimson & Sirmons
28
ODIF Search Results
© 2006 KBSI
Used by permission
Click histogram to
drill down to
document & section
Source documents
selected for search
STC 2006
Target documents
selected to search
Bimson & Sirmons
Results: best (100) to
worst matches
29
Rave (MS Project Plug-in)
Uses KMF to integrate
schedules in different
formats to feed MS Project
STC 2006
Bimson & Sirmons
30
WaveSearch
Searches databases
and documents
simultaneously
© 2006 Modus Operandi, Inc.
Used by permission
STC 2006
Bimson & Sirmons
31
Browse from instance
data in one class
to related instance data
in another
WaveView
View integrated
instance data
Browse ontology
classes & relations
© 2006 Modus Operandi, Inc.
Used by permission
STC 2006
Bimson & Sirmons
32
WaveConsole
View
models
View data
services
Generate
test
queries
View
Results
© 2006 Modus Operandi, Inc.
Used by permission
STC 2006
Bimson & Sirmons
33
Development Process Overview*:
Deploying Ontology-Based EII Solutions
STC 2006
* Reference: Wave
BimsonGetting
& Sirmons Started Guide, Modus Operandi, Inc.34
Wave Ontology-Based EII Development*
Create an ontology (OWL)
Use ontology to
access distributed
information via
web services
Map to data
& document
sources
Revise model
& mappings
STC 2006
Bimson
& Sirmons
* Reference: Wave Getting
Started
Guide, Modus Operandi, Inc.
35
Wave Ontology-Based EII Development
Modify ontology
Facility
contains
Equipment
containedIn
Use ontology to access
NEW & OLD distributed
information via
web services
Add, delete
& modify
Information sources
Revise again
STC 2006
Bimson & Sirmons
36
Wave Ontology-Based EII Development Steps
Ontology Engineering
A. Create OWL ontology in Protégé
B. Create a BEA Workshop application for the ontology
C. Import ontology into BEA AquaLogic
D. Create physical data services
E. Map relationship functions and deploy
Ontology EII
Deployment
F. Configure the Wave Platform
G. Verify that the ontology is deployed
H. Search ontology using Wave Search application
I. Make changes to ontology and re-deploy
STC 2006
Bimson & Sirmons
37
Wave Ontology-Based EII Development Steps
KMF uses Wave/BEA for mapping ontologies
to legacy data and deploying EII solution
Ontology \life cycle management process is
iterative and complex (ref: last section)
Multiple tools: Protégé, BEA, Wave
Each step has multiple sub-steps
Round-trip engineering increases complexity
Experience and training is needed
Ranger & ODIF projects are simplifying OLCM
with better round-trip engineering tools
STC 2006
Bimson & Sirmons
38
Beyond Ontology Engineering:
Ontology Life Cycle Management
STC 2006
Bimson & Sirmons
39
Hard Ontology Life Cycle Questions
Ontology
Maintenance
How do I manage
ontology
evolution, add
new data, and
maintain
consistency?
Exchange
Standard
Generation
Data
Discovery
What data do I
need to integrate
and what does it
look like?
How do I generate
an XML or
standards-based
representation for
other tools to use?
Web
Deployment
How do I deploy
the integrated
data in a useful
way over the
web?
Ontology Design
& Generation
Where do I start
in designing the
ontology…can I
automatically
generate one?
Can I compare
the data
between
databases to see
what fits the
ontology best?
How do I verify
that my ontology
reflects the data?
Where are the
inconsistencies?
How do I map
these databases
to the ontology
and persist the
mappings?
Comparative
Analysis
Database-toModel Mapping &
Persistence
Model
Verification
STC 2006
Bimson & Sirmons
40
OLCM Process Address Hard Questions
Ontology
Maintenance
Exchange
Standard
Generation
Web
Deployment
Data
Discovery
Ontology Design
& Generation
Ontology Life Cycle
Management (OLCM):
Modeling Process
Comparative
Analysis
Treats ontology development as a life
cycle management challenge
Helps modelers manage the
Database-tocomplexities of OLCM
Model Mapping &
Focuses on answering the hard
Persistence
questions
Model
Verification
STC 2006
Bimson & Sirmons
41
Roadmap:
The Way Ahead
STC 2006
Bimson & Sirmons
42
The Way Ahead
Operationalize the use of KMF and SIRP applications
Extend the ontology to new data sources
Develop new SIRP applications
Partner with AFRL, MDA, the 45th Space Wing for
additional SBIR awards and technologies
Extend Space Launch & Range Operations COI
Partner with other companies for new technologies
Provide technology training
STC 2006
Bimson & Sirmons
43