Semantic Web Applications for National Security

1
Putting Context to Work Semantic Keys to Improve
Rapid First Response

• Broadstrokes Group

2
Putting Context to Work Semantic Keys to Improve
Rapid First Response

• Train Derailment Example
• 6 January 2005, 350 A.M. Graniteville, SC
• Chlorine Tank Car Toxic Release
• Matching ResponseType with EventType
• Using Open Public Standards
• Getting the Right Information to the Right People
  at the Right Time

3
Original Timeline v. New Timeline

• Original Timeline 17 Days
• Incident Starts 200 A.M. 6 January 2005
• All Clear 21 January 2005
• New Timeline 12 Days
• Incident Starts 200 A.M. 6 January 2006
• All Clear 17 January 2006

4
Original Timeline Problems

• Incident Starts 200 A.M. 6 January 2005
• No Early Public Warning Alert
• First EPA Situation Report 1000 A.M.
• 8 Hour Response Quick, Given Circumstances
• Injuries, Deaths Acknowledged, Chlorine Hazmat
  Identified
• Second EPA Situation Report 400 P.M.
• EPA Region 4 Operations Support Command (OSC)
• Operations Underway

5
Original Timeline Problems

• Third EPA Situation Report 7 January 2005
• EPA Region 4 OSC on Scene - Air Monitoring
• Extent Hazmat Release Determined 76 of 131 Tons
• Fourth EPA Situation Report 7 January 2005 - Full
  Operations Begin
• Logistical Support Operations Non-NIMS
  Organizational Impedance
• All Clear 17 January 2005

6
New Timeline Improvements

• Incident Starts 200 A.M. 6 January 2005
• COMMON ALERTING PROTOCOL (CAP) Message with
  EMERGENCY DATA EXCHAGE LANGUAGE (EDXL) Header
  Optimizes Immediate Public Warning Alert 215
  A.M.
• Remote Sensors from Tank Cars Automatically
  Trigger Alerts, Trigger Train-Mounted Sirens/PA
  Systems
• First EPA Situation Report 400 A.M.
• Less Than Two-Hour Severe Emergency Response
  Triggered by Fatal-Level Hazmat Release

7
New Timeline Improvements

• Second EPA Situation Report 900 A.M.
• Semantically Triggered Response Allows Full
  Operations to Start within 8 Hours Versus Next
  Day
• Hazmat Automatically Identified, Severity
  Semantics Push First Responses and Eliminate
  Jurisdictional Duplication of Incident Command
  Systems (ICS)
• Logistical Support Operations NIMS Plus
  Semantics Eliminates Organizational Impedance
• All Clear 12 January 2006

8
Semantic Keys to Improvement

• Common Alerting Protocol (CAP) 1.1 Clarification
• Message ID
• CBRN added to Categories
• Chemical, Biological, Radiological, Nuclear
• Digital Signature Encryption
• Emergency Data Exchange Language (EDXL)
• Event Type and Sender/Recipient Type Ontologies
  will Drive Improved Routing

9
Event Type Ontology
Application Protégé http//www.protégé.stanford.e
du
10
Event Type Ontology in Context
Application Unicorn Workbench
http//www.unicorn.com
11
Ontological (Semantic) AnalysisElement Data
Asset Comparison
Result No Duplicated Elements
12
Anatomy of CAP-EDXL Emergency Response
Communications for New Timeline
13
Explanation Point by Point of Diagrams

• 1) Upon Notification of the incident, Sheriff
  calls in using his phone or PDA to trigger an
  emergency alert.
• 2) A CAP message is generated and posted to the
  ICS hub where it is posted as Currently Active
  in message repository regularly polled by
  Emergency Management Community. Rules-Based
  Registry Standards allows a publish-subscribe
  mechanism for propagating alert naming messages.
• 3) NIMS ICS Policy-Based messages are
  automatically routed to the Local EM Agency
  where operators can monitor and send out message
  updates.
• 4) The Secondary Providers Services switch to
  Active Polling for any immediate update-CAP
  messages routed by EDXL Distribution Header
  Component with a PublicAlerting value.
• 5) GIS map tool provider services respond to the
  CAP message and process the information, adding a
  new WSRP Portlet for this Incident. ICS
  incorporates mapping service.
• 6) A reverse 911 geo-coded lookup is performed
  and request for local area code-based automatic
  telephone notification is processes and
  performed.

14
Anatomy of CAP-EDXL Emergency Response
Communications for New Timeline
15
Semantic Web Connects Web Services
16
Adding Value with Web Services
17
Broadstrokes-Reverse 911

• Wireless/VoIP outbound and inbound voice services
• Combines TTS and ASR technologies
• Support for web services, SOAP, VoiceXML, CCXML,
  CAP, XQUERY and XPATH

18
Broadstrokes-Features

• Message creation using Text-to-speech
• Message creation using Voice Recording
• Definition of call lists by static lists
• Answering machine recognition
• Call rules support
• Call Escalation
• Call Bandwidth control
• Critical response collection reporting
• Call Status Reporting
• Multiple Language Options

19
Broadstrokes-WSRP
20
MyStateUSA (Acting as National Incident
Management System-NIMS Incident Command
System-ICS)
21
MyStateUSA
22
TargusInfo

• Alert the public in any given location reach
  more people with the most complete, up-to-date
  name, address, and telephone records.
• Identify unique, non-public resident information
  for emergency use only TARGUSinfo can access
  and deliver telephone data that is not available
  through other public sources.
• Count on data availability during the most
  critical situations our sophisticated network
  has fully redundant systems, ensuring 24/7
  availability.
• Inform the community of non-emergency activities
  current name and address means you can stay
  connected through continuity of operations (COOP)
  announcements, procedures, special assistance and
  new developments.

23
ImageMatters - alertSmarts Features

• Provides the ability to dynamically and a-priori
  geocode addresses using a variety of geocoding
  engines, and then store these as Contacts DB
  through a fast, light-weight indexing engine
  capable of quickly querying millions of
  locations.
• alertSmarts leverages cross-platform Java and
  J2EE technology for extreme scalability to large
  and small system for all levels of government,
  from Federal to local.
• alertSmarts implements standard map server
  interfaces to leverage best-of-breed GIS map
  capabilities. Additionally, alertSmarts
  implements a hardware-accelerated rendering
  engine for alert and call-status information
  which is also available via standard map server
  interfaces.

24
ImageMatters - Features contd

• alertSmarts integrates advanced weather
  capabilities including real-time alerts,
  forecasts and imagery from both NOAA as well as
  the nations largest private weather network with
  over 8000 weather stations nationwide.
• alertSmarts provides advanced location-based
  services for mobile devices supporting the J2ME
  platform. These include mobile alerts, maps and
  status reports.
• Provides the ability to cancel a notification and
  send follow-up notifications to only those
  citizens that were successfully notified.

25
ImageMatters - Geospatial Knowledge Base

• OFC Ontology Foundation Classes
• Support for RDF/OWL
• XPath implementation.
• JDBC
• Driver and DataSource implementations using J2EE
  Connector Architecture.
• javax.sql.RowSet Implementation.
• SQL-based query language.
• Support for OGC Filter Specification, including
  spatial predicates.
• Support for XPath queries
• Stored procedures using Functor framework.

• Resources
• Any mime-typed object.
• Special treatment of RDF/OWL resources.
• Support mime-type transformation using Java
  Activation Framework.
• Metadata based on Dublin Core.
• Folder Model
• Folders are container of resources.
• Folder have a default RDF/OWL resource, much like
  an index.html

26
Starbourne
27
Starbourne